This resource is using open-source code maintained in github (see the quick-start-guide section) and available for download from NGC

The ResNet50 v1.5 model is a modified version of the original ResNet50 v1 model.

The difference between v1 and v1.5 is in the bottleneck blocks which requires\ndownsampling, for example, v1 has stride = 2 in the first 1x1 convolution, whereas v1.5 has stride = 2 in the 3x3 convolution.

This difference makes ResNet50 v1.5 slightly more accurate (~0.5% top1) than v1,\nbut comes with a small performance drawback (~5% imgs/sec).

The following performance optimizations were implemented in this model:

JIT graph compilation with XLA
Multi-GPU training with Horovod
Automated mixed precision AMP

This model is trained with mixed precision using Tensor Cores on Volta, Turing, and the NVIDIA Ampere GPU architectures. Therefore, researchers can get results 3x faster than training without Tensor Cores, while experiencing the benefits of mixed precision training. This model is tested against each NGC monthly container release to ensure consistent accuracy and performance over time.

Default configuration

The following sections highlight the default configuration for the ResNet50 model.

Optimizer

This model uses the SGD optimizer with the following hyperparameters:

Momentum (0.875).
Learning rate (LR) = 0.256 for 256 batch size, for other batch sizes we linearly scale the learning\nrate.
Learning rate schedule - we use cosine LR schedule.
For bigger batch sizes (512 and up) we use linear warmup of the learning rate.\nduring the first 5 epochs according to Training ImageNet in 1 hour.
Weight decay: 3.0517578125e-05 (1/32768).
We do not apply Weight decay on batch norm trainable parameters (gamma/bias).
Label Smoothing: 0.1.
We train for:\n
- 90 Epochs -> 90 epochs is a standard for ResNet family networks.
- 250 Epochs -> best possible accuracy.
\n
For 250 epoch training we also use MixUp regularization.

Data Augmentation

This model uses the following data augmentation:

For training:\n
- Normalization.
- Random resized crop to 224x224.\n
  - Scale from 8% to 100%.
  - Aspect ratio from 3/4 to 4/3.
  \n
- Random horizontal flip.
\n
For inference:\n
- Normalization.
- Scale to 256x256.
- Center crop to 224x224.
\n

Feature support matrix

The following features are supported by this model.

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n

tr:nth-child(odd)]:bg-surface-sunken\">

Feature	ResNet-50 v1.5 Tensorflow
Multi-GPU training with Horovod	Yes
NVIDIA DALI	Yes
Automatic mixed precision (AMP)	Yes

Features

Multi-GPU training with Horovod - Our model uses Horovod to implement efficient multi-GPU training with NCCL.\nFor details, refer to the example sources in this repository or the TensorFlow tutorial.

NVIDIA DALI - DALI is a library accelerating data preparation pipeline. To accelerate your input pipeline, you only need to define your data loader\nwith the DALI library. For details, refer to the example sources in this repository or the DALI documentation.

Automatic mixed precision (AMP) - Computation graph can be modified by TensorFlow on runtime to support mixed precision training.\nDetailed explanation of mixed precision can be found in the next section.

Mixed precision training

Mixed precision is the combined use of different numerical precisions in a computational method. Mixed precision training offers significant computational speedup by performing operations in half-precision format while storing minimal information in single-precision to retain as much information as possible in critical parts of the network. Since the introduction of Tensor Cores in Volta, and following with both the Turing and Ampere architectures, significant training speedups are experienced by switching to mixed precision -- up to 3x overall speedup on the most arithmetically intense model architectures. Using mixed precision training previously required two steps:

Porting the model to use the FP16 data type where appropriate.
Adding loss scaling to preserve small gradient values.

This can now be achieved using Automatic Mixed Precision (AMP) for TensorFlow to enable the full mixed precision methodology in your existing TensorFlow model code. AMP enables mixed precision training on Volta and Turing GPUs automatically. The TensorFlow framework code makes all necessary model changes internally.

In TF-AMP, the computational graph is optimized to use as few casts as necessary and maximize the use of FP16, and the loss scaling is automatically applied inside of supported optimizers. AMP can be configured to work with the existing tf.contrib loss scaling manager by disabling the AMP scaling with a single environment variable to perform only the automatic mixed-precision optimization. It accomplishes this by automatically rewriting all computation graphs with the necessary operations to enable mixed precision training and automatic loss scaling.

For information about:

How to train using mixed precision, see the Mixed Precision Training paper and Training With Mixed Precision documentation.
Techniques used for mixed precision training, see the Mixed-Precision Training of Deep Neural Networks blog.
How to access and enable AMP for TensorFlow, see Using TF-AMP from the TensorFlow User Guide.

Enabling mixed precision

Mixed precision is enabled in TensorFlow by using the Automatic Mixed Precision (TF-AMP) extension which casts variables to half-precision upon retrieval, while storing variables in single-precision format. Furthermore, to preserve small gradient magnitudes in backpropagation, a loss scaling step must be included when applying gradients. In TensorFlow, loss scaling can be applied statically by using simple multiplication of loss by a constant value or automatically, by TF-AMP. Automatic mixed precision makes all the adjustments internally in TensorFlow, providing two benefits over manual operations. First, programmers need not modify network model code, reducing development and maintenance effort. Second, using AMP maintains forward and backward compatibility with all the APIs for defining and running TensorFlow models.

To enable mixed precision, you can simply add the values to the environmental variables inside your training script:

Enable TF-AMP graph rewrite:

os.environ[\"TF_ENABLE_AUTO_MIXED_PRECISION_GRAPH_REWRITE\"] = \"1\"\n

Enable Automated Mixed Precision:

os.environ['TF_ENABLE_AUTO_MIXED_PRECISION'] = '1'\n\n

Enabling TF32

TensorFloat-32 (TF32) is the new math mode in NVIDIA A100 GPUs for handling the matrix math also called tensor operations. TF32 running on Tensor Cores in A100 GPUs can provide up to 10x speedups compared to single-precision floating-point math (FP32) on Volta GPUs.

TF32 Tensor Cores can speed up networks using FP32, typically with no loss of accuracy. It is more robust than FP16 for models which require high dynamic range for weights or activations.

For more information, refer to the TensorFloat-32 in the A100 GPU Accelerates AI Training, HPC up to 20x blog post.

TF32 is supported in the NVIDIA Ampere GPU architecture and is enabled by default.

","resourceVersions",[231,232],"resourceVersionsTotal","hasMoreVersions","versionMetadata",{"_210":28,"_211":-5,"_52":116,"_113":-5,"_117":212,"_119":-5,"_213":115,"_124":16,"_121":34,"_214":-5,"_120":34,"_215":216,"_217":90,"_218":219,"_220":90,"_221":222,"_223":90,"_224":225,"_226":90,"_227":228,"_229":90,"_122":230},"versionNotFound","isLoggedIn","latestVersionPath","/orgs/nvidia/dle/resources/resnet_tf1/-","-","imageReference","cliDownloadCommand","ngc registry resource download-version nvidia/dle/resnet_tf1:20.12.0","helmFetchCommand","viewAllVersionsPath","/orgs/nvidia/dle/resources/resnet_tf1/-/version-history","accessCheckResult",{"_209":90},"hasRelatedCollections","context",{"_205":206,"_207":208},1789368739222,[],[],[198,199,200,201,202,203,204],{"_48":63,"_50":73,"_52":65,"_54":66,"_56":74,"_58":59},{"_48":63,"_50":72,"_52":146,"_54":-5,"_56":-5,"_58":59},{"_48":63,"_50":70,"_52":65,"_54":66,"_56":71,"_58":59},{"_48":63,"_50":68,"_52":65,"_54":66,"_56":69,"_58":59},{"_48":63,"_50":64,"_52":65,"_54":66,"_56":67,"_58":59},{"_48":60,"_50":61,"_52":53,"_54":55,"_56":62,"_58":59},{"_48":49,"_50":51,"_52":53,"_54":55,"_56":57,"_58":59},"orgName","nvidia","teamName","dle","allowed","versionId","description",["D",1663857576461],"totalSizeInBytes","malwareScanStatus","releaseNotes","### Changelog\n1. March, 2019\n * Initial release\n2. May, 2019\n * Added DALI support\n * Added scripts for DGX-2\n * Added benchmark results for DGX-2 and XLA-enabled DGX-1 and DGX-2.\n3. July, 2019\n * Added Cosine learning rate schedule\n4. August, 2019\n * Added mixup regularization\n * Added T4 benchmarks\n * Improved inference capabilities\n * Added SavedModel export \n5. January, 2020\n * Removed manual checks for dataset paths to facilitate cloud storage solutions\n * Move to a new logging solution\n * Bump base docker image version\n6. March, 2020\n * Code cleanup and refactor\n * Improved training process\n7. June, 2020\n * Added benchmark results for DGX-A100\n * Updated benchmark results for DGX-1, DGX-2 and T4\n * Updated base docker image version\n8. August 2020\n * Updated command line argument names\n * Added support for syntetic dataset with different image size\n9. January, 2022\n * Added barrier at the end of multiprocess run\n### Known issues\nPerformance without XLA enabled is low due to BN + ReLU fusion bug.","hasReleaseNotes","setup","The following section lists the requirements that you need to meet in order to use the ResNet50 v1.5 model.\n\n### Requirements\nThis repository contains Dockerfile which extends the TensorFlow NGC container and encapsulates all dependencies. Aside from these dependencies, ensure you have the following software:\n\n- [NVIDIA Docker](https://github.com/NVIDIA/nvidia-docker)\n- [TensorFlow 20.06-tf1-py3 NGC container](https://ngc.nvidia.com/catalog/containers/nvidia:tensorflow)\n- GPU-based architecture:\n - [NVIDIA Volta](https://www.nvidia.com/en-us/data-center/volta-gpu-architecture/)\n - [NVIDIA Turing](https://www.nvidia.com/en-us/geforce/turing/)\n - [NVIDIA Ampere architecture](https://www.nvidia.com/en-us/data-center/nvidia-ampere-gpu-architecture/)\n\nFor more information about how to get started with NGC containers, see the\nfollowing sections from the NVIDIA GPU Cloud Documentation and the Deep Learning Documentation:\n* [Getting Started Using NVIDIA GPU Cloud](https://docs.nvidia.com/ngc/ngc-getting-started-guide/index.html),\n* [Accessing And Pulling From The NGC container registry](https://docs.nvidia.com/deeplearning/frameworks/user-guide/index.html#accessing_registry),\n* [Running TensorFlow](https://docs.nvidia.com/deeplearning/frameworks/tensorflow-release-notes/running.html#running).\n\nFor those unable to use the [TensorFlow NGC container](https://ngc.nvidia.com/catalog/containers/nvidia:tensorflow) to set up the required environment or create your own container, see the versioned [NVIDIA Container Support Matrix](https://docs.nvidia.com/deeplearning/frameworks/support-matrix/index.html).","hasSetup","quickStartGuide","To train your model using mixed precision or TF32 with Tensor Cores or FP32, perform the following steps using the default parameters of the ResNet-50 v1.5 model on the [ImageNet](http://www.image-net.org/) dataset. For the specifics concerning training and inference, see the [Advanced](#advanced) section.\n\n1. Clone the repository.\n```\ngit clone https://github.com/NVIDIA/DeepLearningExamples\ncd DeepLearningExamples/TensorFlow/Classification/ConvNets\n```\n\n2. Download and preprocess the dataset.\nThe ResNet50 v1.5 script operates on ImageNet 1k, a widely popular image classification dataset from the ILSVRC challenge.\n\n* [Download the images](http://image-net.org/download-images)\n* Extract the training and validation data:\n```bash\nmkdir train && mv ILSVRC2012_img_train.tar train/ && cd train\ntar -xvf ILSVRC2012_img_train.tar && rm -f ILSVRC2012_img_train.tar\nfind . -name \"*.tar\" | while read NAME ; do mkdir -p \"${NAME%.tar}\"; tar -xvf \"${NAME}\" -C \"${NAME%.tar}\"; rm -f \"${NAME}\"; done\ncd ..\nmkdir val && mv ILSVRC2012_img_val.tar val/ && cd val && tar -xvf ILSVRC2012_img_val.tar\n```\n* Preprocess dataset to TFRecord form using [script](https://github.com/tensorflow/models/blob/archive/research/inception/inception/data/build_imagenet_data.py). Additional metadata from [autors repository](https://github.com/tensorflow/models/tree/archive/research/inception/inception/data) might be required.\n\n3. Build the ResNet-50 v1.5 TensorFlow NGC container.\n```bash\ndocker build . -t nvidia_rn50\n```\n\n4. Start an interactive session in the NGC container to run training/inference.\nAfter you build the container image, you can start an interactive CLI session with\n```bash\nnvidia-docker run --rm -it -v :/data/tfrecords --ipc=host nvidia_rn50\n```\n\n5. (Optional) Create index files to use DALI.\nTo allow proper sharding in a multi-GPU environment, DALI has to create index files for the dataset. To create index files, run inside the container:\n```bash\nbash ./utils/dali_index.sh /data/tfrecords \n```\nIndex files can be created once and then reused. It is highly recommended to save them into a persistent location.\n\n6. Start training.\nTo run training for a standard configuration (as described in [Default\nconfiguration](#default-configuration), DGX1V, DGX2V, single GPU, FP16, FP32, 50, 90, and 250 epochs), run\none of the scripts int the `resnet50v1.5/training` directory. Ensure ImageNet is mounted in the\n`/data/tfrecords` directory.\n\nFor example, to train on DGX-1 for 90 epochs using AMP, run: \n\n`bash ./resnet50v1.5/training/DGX1_RN50_AMP_90E.sh /path/to/result /data`\n\nAdditionally, features like DALI data preprocessing or TensorFlow XLA can be enabled with\nfollowing arguments when running those scripts:\n\n`bash ./resnet50v1.5/training/DGX1_RN50_AMP_90E.sh /path/to/result /data --xla --dali`\n\n7. Start validation/evaluation.\nTo evaluate the validation dataset located in `/data/tfrecords`, run `main.py` with\n`--mode=evaluate`. For example:\n\n`python main.py --mode=evaluate --data_dir=/data/tfrecords --batch_size --model_dir\n --results_dir [--xla] [--amp]`\n\nThe optional `--xla` and `--amp` flags control XLA and AMP during evaluation.","hasQuickStartGuide","advanced","The following sections provide greater details of the dataset, running training and inference, and the training results.\n\n### Scripts and sample code\n\nIn the root directory, the most important files are:\n - `main.py`: the script that controls the logic of training and validation of the ResNet-like models\n - `Dockerfile`: Instructions for Docker to build a container with the basic set of dependencies to run ResNet like models for image classification\n - `requirements.txt`: a set of extra Python requirements for running ResNet-like models\n\nThe `model/` directory contains the following modules used to define ResNet family models:\n - `resnet.py`: the definition of ResNet, ResNext, and SE-ResNext model\n - `blocks/conv2d_block.py`: the definition of 2D convolution block\n - `blocks/resnet_bottleneck_block.py`: the definition of ResNet-like bottleneck block\n - `layers/*.py`: definitions of specific layers used in the ResNet-like model\n \nThe `utils/` directory contains the following utility modules:\n - `cmdline_helper.py`: helper module for command line processing\n - `data_utils.py`: module defining input data pipelines\n - `dali_utils.py`: helper module for DALI \n - `image_processing.py`: image processing and data augmentation functions\n - `learning_rate.py`: definition of used learning rate schedule\n - `optimizers.py`: definition of used custom optimizers\n - `hooks/*.py`: definitions of specific hooks allowing logging of training and inference process\n \nThe `runtime/` directory contains the following module that define the mechanics of the training process:\n - `runner.py`: module encapsulating the training, inference and evaluation \n\n### Parameters\n\n#### The `main.py` script\nThe script for training and evaluating the ResNet-50 v1.5 model has a variety of parameters that control these processes.\n\n```\nusage: main.py [-h] [--arch {resnet50,resnext101-32x4d,se-resnext101-32x4d}]\n [--mode {train,train_and_evaluate,evaluate,predict,training_benchmark,inference_benchmark}]\n [--export_dir EXPORT_DIR] [--to_predict TO_PREDICT] \n --batch_size BATCH_SIZE [--num_iter NUM_ITER] \n [--run_iter RUN_ITER] [--iter_unit {epoch,batch}] \n [--warmup_steps WARMUP_STEPS] [--model_dir MODEL_DIR]\n [--results_dir RESULTS_DIR] [--log_filename LOG_FILENAME] \n [--display_every DISPLAY_EVERY] [--seed SEED]\n [--gpu_memory_fraction GPU_MEMORY_FRACTION] [--gpu_id GPU_ID]\n [--finetune_checkpoint FINETUNE_CHECKPOINT] [--use_final_conv]\n [--quant_delay QUANT_DELAY] [--quantize] [--use_qdq] \n [--symmetric] [--data_dir DATA_DIR] \n [--data_idx_dir DATA_IDX_DIR] [--dali]\n [--synthetic_data_size SYNTHETIC_DATA_SIZE] [--lr_init LR_INIT]\n [--lr_warmup_epochs LR_WARMUP_EPOCHS] \n [--weight_decay WEIGHT_DECAY] [--weight_init {fan_in,fan_out}]\n [--momentum MOMENTUM] [--label_smoothing LABEL_SMOOTHING]\n [--mixup MIXUP] [--cosine_lr] [--xla] \n [--data_format {NHWC,NCHW}] [--amp]\n [--static_loss_scale STATIC_LOSS_SCALE]\n \nJoC-RN50v1.5-TF \n \noptional arguments: \n -h, --help show this help message and exit.\n --arch {resnet50,resnext101-32x4d,se-resnext101-32x4d}\n Architecture of model to run. \n --mode {train,train_and_evaluate,evaluate,predict,training_benchmark,inference_benchmark}\n The execution mode of the script.\n --export_dir EXPORT_DIR \n Directory in which to write exported SavedModel. \n --to_predict TO_PREDICT \n Path to file or directory of files to run prediction\n on.\n --batch_size BATCH_SIZE \n Size of each minibatch per GPU. \n --num_iter NUM_ITER Number of iterations to run.\n --run_iter RUN_ITER Number of training iterations to run on single run.\n --iter_unit {epoch,batch} \n Unit of iterations. \n --warmup_steps WARMUP_STEPS \n Number of steps considered as warmup and not taken\n into account for performance measurements. \n --model_dir MODEL_DIR \n Directory in which to write model. If undefined, \n results dir will be used. \n --results_dir RESULTS_DIR\n Directory in which to write training logs, summaries\n and checkpoints.\n --log_filename LOG_FILENAME\n Name of the JSON file to which write the training log.\n --display_every DISPLAY_EVERY\n How often (in batches) to print out running\n information.\n --seed SEED Random seed.\n --gpu_memory_fraction GPU_MEMORY_FRACTION\n Limit memory fraction used by training script for DALI.\n --gpu_id GPU_ID Specify ID of the target GPU on multi-device platform.\n Effective only for single-GPU mode.\n --finetune_checkpoint FINETUNE_CHECKPOINT\n Path to pre-trained checkpoint which will be used for\n fine-tuning.\n --use_final_conv Use convolution operator instead of MLP as last layer.\n --quant_delay QUANT_DELAY\n Number of steps to be run before quantization starts\n to happen.\n --quantize Quantize weights and activations during training.\n (Defaults to Assymmetric quantization)\n --use_qdq Use QDQV3 op instead of FakeQuantWithMinMaxVars op for\n quantization. QDQv3 does only scaling.\n --symmetric Quantize weights and activations during training using\n symmetric quantization.\n\nDataset arguments:\n --data_dir DATA_DIR Path to dataset in TFRecord format. Files should be\n named 'train-*' and 'validation-*'.\n --data_idx_dir DATA_IDX_DIR\n Path to index files for DALI. Files should be named\n 'train-*' and 'validation-*'.\n --dali Enable DALI data input.\n --synthetic_data_size SYNTHETIC_DATA_SIZE\n Dimension of image for synthetic dataset.\n\nTraining arguments:\n --lr_init LR_INIT Initial value for the learning rate.\n --lr_warmup_epochs LR_WARMUP_EPOCHS\n Number of warmup epochs for learning rate schedule.\n --weight_decay WEIGHT_DECAY\n Weight Decay scale factor.\n --weight_init {fan_in,fan_out}\n Model weight initialization method.\n --momentum MOMENTUM SGD momentum value for the Momentum optimizer.\n --label_smoothing LABEL_SMOOTHING\n The value of label smoothing.\n --mixup MIXUP The alpha parameter for mixup (if 0 then mixup is not\n applied).\n --cosine_lr Use cosine learning rate schedule.\n\nGeneric optimization arguments:\n --xla Enable XLA (Accelerated Linear Algebra) computation\n for improved performance.\n --data_format {NHWC,NCHW}\n Data format used to do calculations.\n --amp Enable Automatic Mixed Precision to speedup\n computation using tensor cores.\n\nAutomatic Mixed Precision arguments:\n --static_loss_scale STATIC_LOSS_SCALE\n Use static loss scaling in FP32 AMP.\n\n```\n\n### Quantization Aware Training\nQuantization Aware training (QAT) simulates quantization during training by quantizing weights and activation layers. This will help reduce the loss in accuracy when we convert the network\ntrained in FP32 to INT8 for faster inference. QAT introduces additional nodes in the graph which will be used to learn the dynamic ranges of weights and activation layers. Tensorflow provides\na quantization tool which automatically adds these nodes in-place. Typical workflow\nfor training QAT networks is to train a model until convergence and then finetune with the quantization layers. It is recommended that QAT is performed on a single GPU.\n\n* For 1 GPU\n * Command: `sh resnet50v1.5/training/GPU1_RN50_QAT.sh `\n \nIt is recommended to finetune a model with quantization nodes rather than train a QAT model from scratch. The latter can also be performed by setting `quant_delay` parameter.\n`quant_delay` is the number of steps after which quantization nodes are added for QAT. If we are fine-tuning, `quant_delay` is set to 0. \n \nFor QAT network, we use tf.quantization.quantize_and_dequantize operation.\nThese operations are automatically added at weights and activation layers in the RN50 by using `tf.contrib.quantize.experimental_create_training_graph` utility. Support for using `tf.quantization.quantize_and_dequantize` \noperations for `tf.contrib.quantize.experimental_create_training_graph` has been added in TensorFlow 20.01-py3 NGC container and later versions, which is required for this task.\n\n#### Post process checkpoint\n `postprocess_ckpt.py` is a utility to convert the final classification FC layer into a 1x1 convolution layer using the same weights. This is required to ensure TensorRT can parse QAT models successfully.\n This script should be used after performing QAT to reshape the FC layer weights in the final checkpoint.\n Arguments:\n * `--input` : Path to the trained checkpoint of RN50.\n * `--output` : Name of the new checkpoint file which has the FC layer weights reshaped into 1x1 conv layer weights.\n * `--dense_layer` : Name of the FC layer\n\n#### Exporting Frozen graphs\nTo export frozen graphs (which can be used for inference with TensorRT), use:\n\n`python export_frozen_graph.py --checkpoint --quantize --use_final_conv --use_qdq --symmetric --input_format NCHW --compute_format NCHW --output_file=`\n\nArguments:\n\n* `--checkpoint` : Optional argument to export the model with checkpoint weights.\n* `--quantize` : Optional flag to export quantized graphs.\n* `--use_qdq` : Use quantize_and_dequantize (QDQ) op instead of FakeQuantWithMinMaxVars op for quantization. QDQ does only scaling. \n* `--input_format` : Data format of input tensor (Default: NCHW). Use NCHW format to optimize the graph with TensorRT.\n* `--compute_format` : Data format of the operations in the network (Default: NCHW). Use NCHW format to optimize the graph with TensorRT.\n\n### Inference process\nTo run inference on a single example with a checkpoint and a model script, use: \n\n`python main.py --mode predict --model_dir --to_predict --results_dir `\n\nThe optional `--xla` and `--amp` flags control XLA and AMP during inference. To run inference using TF-TRT, please use the following command:\n\n`python inference.py --model --tf-trt --batch-size --data-dir `\n\nThe optional `--amp` flag controls AMP during inference.","hasAdvanced","performance","The performance measurements in this document were conducted at the time of publication and may not reflect the performance achieved from NVIDIA's latest software release. For the most up-to-date performance measurements, go to [NVIDIA Data Center Deep Learning Product Performance](https://developer.nvidia.com/deep-learning-performance-training-inference).\n\n### Benchmarking\n\nThe following section shows how to run benchmarks measuring the model performance in training and inference modes.\n\n#### Training performance benchmark\n\nTo benchmark the training performance on a specific batch size, run:\n\n* For 1 GPU\n * FP32 / TF32\n\n `python ./main.py --mode=training_benchmark --warmup_steps 200 --batch_size --data_dir= --results_dir=`\n \n * AMP\n\n `python ./main.py --mode=training_benchmark --amp --warmup_steps 200 --batch_size --data_dir= --results_dir=`\n \n* For multiple GPUs\n * FP32 / TF32\n\n `mpiexec --allow-run-as-root --bind-to socket -np python ./main.py --mode=training_benchmark --batch_size --data_dir= --results_dir=`\n \n * AMP\n \n `mpiexec --allow-run-as-root --bind-to socket -np python ./main.py --mode=training_benchmark --amp --batch_size --data_dir= --results_dir=`\n \nEach of these scripts runs 200 warm-up iterations and measures the first epoch.\n\nTo control warmup and benchmark length, use the `--warmup_steps`, `--num_iter` and `--iter_unit` flags. Features like XLA or DALI can be controlled\nwith `--xla` and `--dali` flags. For proper throughput reporting the value of `--num_iter` must be greater than `--warmup_steps` value.\nSuggested batch sizes for training are 256 for mixed precision training and 128 for single precision training per single V100 16 GB.\n\nIf no `--data_dir=` flag is specified then the benchmarks will use a synthetic dataset. The resolution of synthetic images used can be controlled with `--synthetic_data_size` flag.\n\n#### Inference performance benchmark\n\nTo benchmark the inference performance on a specific batch size, run:\n\n* FP32 / TF32\n\n`python ./main.py --mode=inference_benchmark --warmup_steps 20 --num_iter 100 --iter_unit batch --batch_size --data_dir= --results_dir=`\n\n* AMP\n\n`python ./main.py --mode=inference_benchmark --amp --warmup_steps 20 --num_iter 100 --iter_unit batch --batch_size --data_dir= --results_dir=`\n\nBy default, each of these scripts runs 20 warm-up iterations and measures the next 80 iterations.\nTo control warm-up and benchmark length, use the `--warmup_steps`, `--num_iter` and `--iter_unit` flags. \nIf no `--data_dir=` flag is specified then the benchmarks will use a synthetic dataset.\n\nThe benchmark can be automated with the `inference_benchmark.sh` script provided in `resnet50v1.5`, by simply running:\n`bash ./resnet50v1.5/inference_benchmark.sh `\n\nThe `` parameter refers to the input data directory (by default `/data/tfrecords` inside the container). \nBy default, the benchmark tests the following configurations: **FP32**, **AMP**, **AMP + XLA** with different batch sizes.\nWhen the optional directory with the DALI index files `` is specified, the benchmark executes an additional **DALI + AMP + XLA** configuration.\nFor proper throughput reporting the value of `--num_iter` must be greater than `--warmup_steps` value.\n\nFor performance benchmark of raw model, synthetic dataset can be used. To use synthetic dataset, use `--synthetic_data_size` flag instead of `--data_dir` to specify input image size.\n\n### Results\n\nThe following sections provide details on how we achieved our performance and accuracy in training and inference. \n\n#### Training accuracy results\n\n##### Training accuracy: NVIDIA DGX A100 (8x A100 40GB)\n\nOur results were obtained by running the `/resnet50v1.5/training/DGXA100_RN50_{PRECISION}_90E.sh` \ntraining script in the [TensorFlow 20.06-tf1-py3 NGC container](https://ngc.nvidia.com/catalog/containers/nvidia:tensorflow) \nNGC container on NVIDIA DGX A100 (8x A100 40GB) GPUs.\n\n| Epochs | Batch Size / GPU | Accuracy - TF32 (top1) | Accuracy - mixed precision (top1) | \n|--------|------------------|-----------------|----------------------------|\n| 90 | 256 | 77.01 | 76.93 |\n\n##### Training accuracy: NVIDIA DGX-1 (8x V100 16G)\nOur results were obtained by running the `/resnet50v1.5/training/DGX1_RN50_{PRECISION}_{EPOCHS}E.sh` \ntraining script in the [TensorFlow 20.06-tf1-py3 NGC container](https://ngc.nvidia.com/catalog/containers/nvidia:tensorflow) \nNGC container on NVIDIA DGX-1 with (8x V100 16G) GPUs.\n\n| Epochs | Batch Size / GPU | Accuracy - FP32 | Accuracy - mixed precision | \n|--------|------------------|-----------------|----------------------------|\n| 90 | 128 (FP32) / 256 (AMP) | 77.01 | 76.99 |\n| 250 | 128 (FP32) / 256 (AMP) | 78.34 | 78.35 |\n\n**Example training loss plot**\n\n![TrainingLoss](https://raw.githubusercontent.com/NVIDIA/DeepLearningExamples/master/TensorFlow/Classification/ConvNets/resnet50v1.5//imgs/train_loss.png)\n\n\n#### Training performance results\n\n##### Training performance: NVIDIA DGX A100 (8x A100 40GB)\nOur results were obtained by running the `resnet50v1.5/training/training_perf.sh` benchmark script in the \n[TensorFlow 20.06-tf1-py3 NGC container](https://ngc.nvidia.com/catalog/containers/nvidia:tensorflow) NGC container \non NVIDIA DGX A100 (8x A100 40GB) GPUs. Performance numbers (in images per second) were averaged over an entire training epoch.\n\n| GPUs | Batch Size / GPU | Throughput - TF32 + XLA | Throughput - mixed precision + XLA | Throughput speedup (TF32 - mixed precision) | Weak scaling - TF32 + XLA | Weak scaling - mixed precision + XLA |\n|----|---------------|---------------|------------------------|-----------------|-----------|-------------------|\n| 1 | 256 | 909 img/s | 2375 img/s | 2.60x | 1.00x | 1.00x |\n| 8 | 256 | 7000 img/s | 17400 img/s | 2.48x | 7.70x | 7.32x |\n\n##### Training performance: NVIDIA DGX-1 (8x V100 16G)\nOur results were obtained by running the `resnet50v1.5/training/training_perf.sh` benchmark script in the \n[TensorFlow 20.06-tf1-py3 NGC container](https://ngc.nvidia.com/catalog/containers/nvidia:tensorflow) NGC container \non NVIDIA DGX-1 with (8x V100 16G) GPUs. Performance numbers (in images per second) were averaged over an entire training epoch.\n\n| GPUs | Batch Size / GPU | Throughput - FP32 + XLA | Throughput - mixed precision + XLA | Throughput speedup (FP32 - mixed precision) | Weak scaling - FP32 + XLA | Weak scaling - mixed precision + XLA |\n|----|---------------|---------------|------------------------|-----------------|-----------|-------------------|\n| 1 | 128 (FP32) / 256 (AMP) | 412 img/s | 1270 img/s | 3.08x | 1.00x | 1.00x |\n| 8 | 128 (FP32) / 256 (AMP) | 3170 img/s | 9510 img/s | 3.00x | 7.69x | 7.48x |\n\n##### Training performance: NVIDIA DGX-2 (16x V100 32G)\nOur results were obtained by running the `resnet50v1.5/training/training_perf.sh` benchmark script in the \n[TensorFlow 20.06-tf1-py3 NGC container](https://ngc.nvidia.com/catalog/containers/nvidia:tensorflow) NGC container \non NVIDIA DGX-2 with (16x V100 32G) GPUs. Performance numbers (in images per second) were averaged over an entire training epoch.\n\n| GPUs | Batch Size / GPU | Throughput - FP32 + XLA | Throughput - mixed precision + XLA | Throughput speedup (FP32 - mixed precision) | Weak scaling - FP32 + XLA | Weak scaling - mixed precision + XLA |\n|----|---------------|---------------|-------------------------|-------|--------|--------|\n| 1 | 128 (FP32) / 256 (AMP) | 432 img/s | 1300 img/s | 3.01x | 1.00x | 1.00x |\n| 16 | 128 (FP32) / 256 (AMP) | 6500 img/s | 17250 img/s | 2.65x | 15.05x | 13.27x |\n\n#### Training Time for 90 Epochs\n\n##### Training time: NVIDIA DGX A100 (8x A100 40GB)\n\nOur results were estimated based on the [training performance results](#training-performance-nvidia-dgx-a100-8x-a100-40g) \non NVIDIA DGX A100 with (8x A100 40G) GPUs.\n\n| GPUs | Time to train - mixed precision + XLA | Time to train - TF32 + XLA |\n|---|--------|---------|\n| 1 | ~18h | ~40h |\n| 8 | ~2h | ~5h | \n\n##### Training time: NVIDIA DGX-1 (8x V100 16G)\n\nOur results were estimated based on the [training performance results](#training-performance-nvidia-dgx-1-8x-v100-16g) \non NVIDIA DGX-1 with (8x V100 16G) GPUs.\n\n| GPUs | Time to train - mixed precision + XLA | Time to train - FP32 + XLA |\n|---|--------|---------|\n| 1 | ~25h | ~77h |\n| 8 | ~3.5h | ~10h | \n\n##### Training time: NVIDIA DGX-2 (16x V100 32G)\n\nOur results were estimated based on the [training performance results](#training-performance-nvidia-dgx-2-16x-v100-32g) \non NVIDIA DGX-2 with (16x V100 32G) GPUs.\n\n| GPUs | Time to train - mixed precision + XLA | Time to train - FP32 + XLA |\n|----|-------|--------|\n| 1 | ~25h | ~74h |\n| 16 | ~2h | ~5h | \n\n#### Inference performance results\n\n##### Inference performance: NVIDIA DGX A100 (1x A100 40GB)\n\nOur results were obtained by running the `inference_benchmark.sh` inferencing benchmarking script\nin the [TensorFlow 20.06-tf1-py3 NGC container](https://ngc.nvidia.com/catalog/containers/nvidia:tensorflow) NGC container \non NVIDIA DGX A100 with (1x A100 40G) GPU.\n\n**TF32 Inference Latency**\n\n|**Batch Size**|**Avg throughput**|**Avg latency**|**90% Latency**|**95% Latency**|**99% Latency**|\n|--------------|------------------|---------------|---------------|---------------|---------------|\n| 1 | 191.23 img/s | 5.26 ms | 5.29 ms | 5.31 ms | 5.42 ms |\n| 2 | 376.83 img/s | 5.34 ms | 5.36 ms | 5.39 ms | 5.56 ms |\n| 4 | 601.12 img/s | 6.65 ms | 6.80 ms | 6.93 ms | 7.05 ms |\n| 8 | 963.86 img/s | 8.31 ms | 8.63 ms | 8.80 ms | 9.17 ms |\n| 16 | 1361.58 img/s | 11.82 ms | 12.04 ms | 12.15 ms | 12.44 ms |\n| 32 | 1602.09 img/s | 19.99 ms | 20.48 ms | 20.74 ms | 21.36 ms |\n| 64 | 1793.81 img/s | 35.82 ms | 37.22 ms | 37.43 ms | 37.84 ms |\n| 128 | 1876.22 img/s | 68.23 ms | 69.60 ms | 70.08 ms | 70.70 ms |\n| 256 | 1911.96 img/s | 133.90 ms | 135.16 ms | 135.59 ms | 136.49 ms |\n\n**TF32 Inference Latency + XLA**\n\n|**Batch Size**|**Avg throughput**|**Avg latency**|**90% Latency**|**95% Latency**|**99% Latency**|\n|--------------|------------------|---------------|---------------|---------------|---------------|\n| 1 | 158.67 img/s | 6.34 ms | 6.39 ms | 6.46 ms | 7.16 ms |\n| 2 | 321.83 img/s | 6.24 ms | 6.29 ms | 6.34 ms | 6.39 ms |\n| 4 | 574.28 img/s | 7.01 ms | 7.03 ms | 7.06 ms | 7.14 ms |\n| 8 | 1021.20 img/s | 7.84 ms | 8.00 ms | 8.08 ms | 8.28 ms |\n| 16 | 1515.79 img/s | 10.56 ms | 10.88 ms | 10.98 ms | 11.22 ms |\n| 32 | 1945.44 img/s | 16.46 ms | 16.78 ms | 16.96 ms | 17.49 ms |\n| 64 | 2313.13 img/s | 27.81 ms | 28.68 ms | 29.10 ms | 30.33 ms |\n| 128 | 2449.88 img/s | 52.27 ms | 54.00 ms | 54.43 ms | 56.85 ms |\n| 256 | 2548.87 img/s | 100.45 ms | 102.34 ms | 103.04 ms | 104.81 ms |\n\n**Mixed Precision Inference Latency**\n\n|**Batch Size**|**Avg throughput**|**Avg latency**|**90% Latency**|**95% Latency**|**99% Latency**|\n|--------------|------------------|---------------|---------------|---------------|---------------|\n| 1 | 223.35 img/s | 4.51 ms | 4.50 ms | 4.52 ms | 4.76 ms |\n| 2 | 435.51 img/s | 4.63 ms | 4.62 ms | 4.64 ms | 4.76 ms |\n| 4 | 882.00 img/s | 4.63 ms | 4.60 ms | 4.71 ms | 5.36 ms |\n| 8 | 1503.24 img/s | 5.40 ms | 5.50 ms | 5.59 ms | 5.78 ms |\n| 16 | 1903.58 img/s | 8.47 ms | 8.67 ms | 8.77 ms | 9.14 ms |\n| 32 | 1974.01 img/s | 16.23 ms | 16.65 ms | 16.96 ms | 17.98 ms |\n| 64 | 3570.46 img/s | 18.14 ms | 18.26 ms | 18.43 ms | 19.35 ms |\n| 128 | 3474.94 img/s | 37.86 ms | 44.09 ms | 55.30 ms | 66.90 ms |\n| 256 | 3229.32 img/s | 81.02 ms | 96.21 ms | 105.67 ms | 126.31 ms |\n\n**Mixed Precision Inference Latency + XLA**\n\n|**Batch Size**|**Avg throughput**|**Avg latency**|**90% Latency**|**95% Latency**|**99% Latency**|\n|--------------|------------------|---------------|---------------|---------------|---------------|\n| 1 | 174.68 img/s | 5.76 ms | 5.81 ms | 5.95 ms | 6.13 ms |\n| 2 | 323.90 img/s | 6.21 ms | 6.26 ms | 6.31 ms | 6.64 ms |\n| 4 | 639.75 img/s | 6.25 ms | 6.45 ms | 6.55 ms | 6.79 ms |\n| 8 | 1215.50 img/s | 6.59 ms | 6.94 ms | 7.03 ms | 7.25 ms |\n| 16 | 2219.96 img/s | 7.29 ms | 7.45 ms | 7.57 ms | 8.09 ms |\n| 32 | 2363.70 img/s | 13.70 ms | 13.91 ms | 14.08 ms | 14.64 ms |\n| 64 | 3940.95 img/s | 18.76 ms | 26.58 ms | 35.41 ms | 59.06 ms |\n| 128 | 3274.01 img/s | 41.70 ms | 52.19 ms | 61.14 ms | 78.68 ms |\n| 256 | 3676.14 img/s | 71.67 ms | 82.36 ms | 88.53 ms | 108.18 ms |\n\n##### Inference performance: NVIDIA DGX-1 (1x V100 16G)\n\nOur results were obtained by running the `inference_benchmark.sh` inferencing benchmarking script\nin the [TensorFlow 20.06-tf1-py3 NGC container](https://ngc.nvidia.com/catalog/containers/nvidia:tensorflow) NGC container \non NVIDIA DGX-1 with (1x V100 16G) GPU.\n\n**FP32 Inference Latency**\n\n|**Batch Size**|**Avg throughput**|**Avg latency**|**90% Latency**|**95% Latency**|**99% Latency**|\n|--------------|------------------|---------------|---------------|---------------|---------------|\n|1\t|173.35 img/s\t|5.79 ms\t|5.90 ms\t|5.95 ms\t|6.04 ms |\n|2\t|303.65 img/s\t|6.61 ms\t|6.80 ms\t|6.87 ms\t|7.01 ms |\n|4\t|562.35 img/s\t|7.12 ms\t|7.32 ms\t|7.42 ms\t|7.69 ms |\n|8\t|783.24 img/s\t|10.22 ms\t|10.37 ms\t|10.44 ms\t|10.60 ms |\n|16\t|1003.10 img/s\t|15.99 ms\t|16.07 ms\t|16.12 ms\t|16.29 ms |\n|32\t|1140.12 img/s\t|28.19 ms\t|28.27 ms\t|28.38 ms\t|28.54 ms |\n|64\t|1252.06 img/s\t|51.12 ms\t|51.82 ms\t|52.75 ms\t|53.45 ms |\n|128\t|1324.91 img/s\t|96.61 ms\t|97.02 ms\t|97.25 ms\t|99.08 ms |\n|256\t|1348.52 img/s\t|189.85 ms\t|191.16 ms\t|191.77 ms\t|192.47 ms|\n\n**Mixed Precision Inference Latency**\n\n|**Batch Size**|**Avg throughput**|**Avg latency**|**90% Latency**|**95% Latency**|**99% Latency**|\n|--------------|------------------|---------------|---------------|---------------|---------------|\n|1\t|237.35 img/s\t|4.25 ms\t|4.39 ms\t|4.54 ms\t|5.30 ms |\n|2\t|464.94 img/s\t|4.32 ms\t|4.63 ms\t|4.83 ms\t|5.52 ms |\n|4\t|942.44 img/s\t|4.26 ms\t|4.55 ms\t|4.74 ms\t|5.45 ms |\n|8\t|1454.93 img/s\t|5.57 ms\t|5.73 ms\t|5.91 ms\t|6.51 ms |\n|16\t|2003.75 img/s\t|8.13 ms\t|8.19 ms\t|8.29 ms\t|8.50 ms |\n|32\t|2356.17 img/s\t|13.69 ms\t|13.82 ms\t|13.92 ms\t|14.26 ms |\n|64\t|2706.11 img/s\t|23.86 ms\t|23.82 ms\t|23.89 ms\t|24.10 ms |\n|128\t|2770.61 img/s\t|47.04 ms\t|49.36 ms\t|62.43 ms\t|90.05 ms |\n|256\t|2742.14 img/s\t|94.67 ms\t|108.02 ms\t|119.34 ms\t|145.55 ms|\n\n**Mixed Precision Inference Latency + XLA**\n\n|**Batch Size**|**Avg throughput**|**Avg latency**|**90% Latency**|**95% Latency**|**99% Latency**|\n|--------------|------------------|---------------|---------------|---------------|---------------|\n|1\t|162.95 img/s\t|6.16 ms\t|6.28 ms\t|6.34 ms\t|6.50 ms |\n|2\t|335.63 img/s\t|5.96 ms\t|6.10 ms\t|6.14 ms\t|6.25 ms |\n|4\t|637.72 img/s\t|6.30 ms\t|6.53 ms\t|7.17 ms\t|8.10 ms |\n|8\t|1153.92 img/s\t|7.03 ms\t|7.97 ms\t|8.22 ms\t|9.00 ms |\n|16\t|1906.52 img/s\t|8.64 ms\t|9.51 ms\t|9.88 ms\t|10.47 ms |\n|32\t|2492.78 img/s\t|12.84 ms\t|13.06 ms\t|13.13 ms\t|13.24 ms |\n|64\t|2910.05 img/s\t|22.66 ms\t|21.82 ms\t|24.71 ms\t|48.61 ms |\n|128\t|2964.31 img/s\t|45.25 ms\t|59.30 ms\t|71.42 ms\t|98.72 ms |\n|256\t|2898.12 img/s\t|90.53 ms\t|106.12 ms\t|118.12 ms\t|150.78 ms|\n\n##### Inference performance: NVIDIA DGX-2 (1x V100 32G)\n\nOur results were obtained by running the `inference_benchmark.sh` inferencing benchmarking script\nin the [TensorFlow 20.06-tf1-py3 NGC container](https://ngc.nvidia.com/catalog/containers/nvidia:tensorflow) NGC container \non NVIDIA DGX-2 with (1x V100 32G) GPU.\n\n**FP32 Inference Latency**\n\n|**Batch Size**|**Avg throughput**|**Avg latency**|**90% Latency**|**95% Latency**|**99% Latency**|\n|--------------|------------------|---------------|---------------|---------------|---------------|\n|1\t|187.41 img/s\t|5.374 ms\t|5.61 ms\t|5.70 ms\t|6.33 ms |\n|2\t|339.52 img/s\t|5.901 ms\t|6.16 ms\t|6.29 ms\t|6.53 ms |\n|4\t|577.50 img/s\t|6.940 ms\t|7.07 ms\t|7.24 ms\t|7.99 ms |\n|8\t|821.15 img/s\t|9.751 ms\t|9.99 ms\t|10.15 ms\t|10.80 ms|\n|16\t|1055.64 img/s\t|15.209 ms\t|15.26 ms\t|15.30 ms\t|16.14 ms|\n|32\t|1195.74 img/s\t|26.772 ms\t|26.93 ms\t|26.98 ms\t|27.80 ms|\n|64\t|1313.83 img/s\t|48.796 ms\t|48.99 ms\t|49.72 ms\t|51.83 ms|\n|128\t|1372.58 img/s\t|93.262 ms\t|93.90 ms\t|94.97 ms\t|96.57 ms|\n|256\t|1414.99 img/s\t|180.923 ms\t|181.65 ms\t|181.92 ms\t|183.37 ms|\n\n**Mixed Precision Inference Latency**\n\n|**Batch Size**|**Avg throughput**|**Avg latency**|**90% Latency**|**95% Latency**|**99% Latency**|\n|--------------|------------------|---------------|---------------|---------------|---------------|\n|1\t|289.89 img/s\t|3.50 ms\t|3.81 ms\t|3.90 ms\t|4.19 ms |\n|2\t|606.27 img/s\t|3.38 ms\t|3.56 ms\t|3.76 ms\t|4.25 ms |\n|4\t|982.92 img/s\t|4.09 ms\t|4.42 ms\t|4.53 ms\t|4.81 ms |\n|8\t|1553.34 img/s\t|5.22 ms\t|5.31 ms\t|5.50 ms\t|6.74 ms |\n|16\t|2091.27 img/s\t|7.82 ms\t|7.77 ms\t|7.82 ms\t|8.77 ms |\n|32\t|2457.61 img/s\t|13.14 ms\t|13.15 ms\t|13.21 ms\t|13.37 ms|\n|64\t|2746.11 img/s\t|23.31 ms\t|23.50 ms\t|23.56 ms\t|24.31 ms|\n|128\t|2937.20 img/s\t|43.58 ms\t|43.76 ms\t|43.82 ms\t|44.37 ms|\n|256\t|3009.83 img/s\t|85.06 ms\t|86.23 ms\t|87.37 ms\t|88.67 ms|\n\n**Mixed Precision Inference Latency + XLA**\n\n|**Batch Size**|**Avg throughput**|**Avg latency**|**90% Latency**|**95% Latency**|**99% Latency**|\n|--------------|------------------|---------------|---------------|---------------|---------------|\n|1\t|240.66 img/s\t|4.22 ms\t|4.59 ms\t|4.69 ms\t|4.84 ms |\n|2\t|428.60 img/s\t|4.70 ms\t|5.11 ms\t|5.44 ms\t|6.01 ms |\n|4\t|945.38 img/s\t|4.26 ms\t|4.35 ms\t|4.42 ms\t|4.74 ms |\n|8\t|1518.66 img/s\t|5.33 ms\t|5.50 ms\t|5.63 ms\t|5.88 ms |\n|16\t|2091.66 img/s\t|7.83 ms\t|7.74 ms\t|7.79 ms\t|8.88 ms |\n|32\t|2604.17 img/s\t|12.40 ms\t|12.45 ms\t|12.51 ms\t|12.61 ms|\n|64\t|3101.15 img/s\t|20.64 ms\t|20.93 ms\t|21.00 ms\t|21.17 ms|\n|128\t|3408.72 img/s\t|37.55 ms\t|37.93 ms\t|38.05 ms\t|38.53 ms|\n|256\t|3633.85 img/s\t|70.85 ms\t|70.93 ms\t|71.12 ms\t|71.45 ms|\n\n##### Inference performance: NVIDIA T4 (1x T4 16G)\n\nOur results were obtained by running the `inference_benchmark.sh` inferencing benchmarking script\nin the [TensorFlow 20.06-tf1-py3 NGC container](https://ngc.nvidia.com/catalog/containers/nvidia:tensorflow) NGC container \non NVIDIA T4 with (1x T4 16G) GPU.\n\n**FP32 Inference Latency**\n\n|**Batch Size**|**Avg throughput**|**Avg latency**|**90% Latency**|**95% Latency**|**99% Latency**|\n|--------------|------------------|---------------|---------------|---------------|---------------|\n|1\t|136.44 img/s\t|7.34 ms\t|7.43 ms\t|7.47 ms\t|7.54 ms |\n|2\t|215.38 img/s\t|9.29 ms\t|9.42 ms\t|9.46 ms\t|9.59 ms |\n|4\t|289.29 img/s\t|13.83 ms\t|14.08 ms\t|14.16 ms\t|14.40 ms |\n|8\t|341.77 img/s\t|23.41 ms\t|23.79 ms\t|23.86 ms\t|24.11 ms |\n|16\t|394.36 img/s\t|40.58 ms\t|40.87 ms\t|40.98 ms\t|41.41 ms |\n|32\t|414.66 img/s\t|77.18 ms\t|78.05 ms\t|78.29 ms\t|78.67 ms |\n|64\t|424.42 img/s\t|150.82 ms\t|152.99 ms\t|153.44 ms\t|154.34 ms |\n|128\t|429.83 img/s\t|297.82 ms\t|301.09 ms\t|301.60 ms\t|302.51 ms |\n|256\t|425.72 img/s\t|601.37 ms\t|605.74 ms\t|606.47 ms\t|608.74 ms |\n\n**Mixed Precision Inference Latency**\n\n|**Batch Size**|**Avg throughput**|**Avg latency**|**90% Latency**|**95% Latency**|**99% Latency**|\n|--------------|------------------|---------------|---------------|---------------|---------------|\n|1\t|211.04 img/s\t|4.77 ms\t|5.05 ms\t|5.08 ms\t|5.15 ms |\n|2\t|381.23 img/s\t|5.27 ms\t|5.40 ms\t|5.45 ms\t|5.52 ms |\n|4\t|593.13 img/s\t|6.75 ms\t|6.89 ms\t|6.956 ms\t|7.02 ms |\n|8\t|791.12 img/s\t|10.16 ms\t|10.35 ms\t|10.43 ms\t|10.68 ms |\n|16\t|914.26 img/s\t|17.55 ms\t|17.80 ms\t|17,89 ms\t|18.19 ms |\n|32\t|972.36 img/s\t|32.92 ms\t|33.33 ms\t|33.46 ms\t|33.61 ms |\n|64\t|991.39 img/s\t|64.56 ms\t|65.62 ms\t|65.92 ms\t|66.35 ms |\n|128\t|995.81 img/s\t|128.55 ms\t|130.03 ms\t|130.37 ms\t|131.08 ms |\n|256\t|993.39 img/s\t|257.71 ms\t|259.26 ms\t|259.62 ms\t|260.36 ms |\n\n**Mixed Precision Inference Latency + XLA**\n\n|**Batch Size**|**Avg throughput**|**Avg latency**|**90% Latency**|**95% Latency**|**99% Latency**|\n|--------------|------------------|---------------|---------------|---------------|---------------|\n|1 |167.01 img/s\t|6.01 ms\t|6.12 ms\t|6.14 ms\t|6.18 ms |\n|2\t|333.67 img/s\t|6.03 ms\t|6.11 ms\t|6.15 ms\t|6.23 ms |\n|4\t|605.94 img/s\t|6.63 ms\t|6.79 ms\t|6.86 ms\t|7.02 ms |\n|8\t|802.13 img/s\t|9.98 ms\t|10.14 ms\t|10.22 ms\t|10.36 ms |\n|16\t|986.85 img/s\t|16.27 ms\t|16.36 ms\t|16.42 ms\t|16.52 ms |\n|32\t|1090.38 img/s\t|29.35 ms\t|29.68 ms\t|29.79 ms\t|30.07 ms |\n|64\t|1131.56 img/s\t|56.63 ms\t|57.22 ms\t|57.41 ms\t|57.76 ms |\n|128\t|1167.62 img/s\t|109.77 ms\t|111.06 ms\t|111.27 ms\t|111.85 ms|\n|256\t|1193.74 img/s\t|214.46 ms\t|216.28 ms\t|216.86 ms\t|217.80 ms|","hasPerformance",[],{"_85":182,"_112":234,"_113":-5,"_121":34,"_120":34,"_122":233,"_217":90},{"_85":28,"_112":28,"_113":-5,"_121":34,"_120":34,"_122":233,"_217":90},[],"Latest (20.12.0)","nvidia/dle/resnet_tf1","key","nvidia/dle/resources/resnet_tf1","resourceId","resnet_tf1","shortDescription","With modified architecture and initialization this ResNet50 version gives ~0.5% better accuracy than original.","This resource is using open-source code maintained in github (see the quick-start-guide section) and available for download from [NGC](https://catalog.ngc.nvidia.com/orgs/nvidia/teams/dle/resources/convnets-tf1)\n\nThe ResNet50 v1.5 model is a modified version of the [original ResNet50 v1 model](https://arxiv.org/abs/1512.03385).\n\nThe difference between v1 and v1.5 is in the bottleneck blocks which requires\ndownsampling, for example, v1 has stride = 2 in the first 1x1 convolution, whereas v1.5 has stride = 2 in the 3x3 convolution.\n\nThis difference makes ResNet50 v1.5 slightly more accurate (\\~0.5% top1) than v1,\nbut comes with a small performance drawback (\\~5% imgs/sec).\n\nThe following performance optimizations were implemented in this model:\n* JIT graph compilation with [XLA](https://www.tensorflow.org/xla)\n* Multi-GPU training with [Horovod](https://github.com/horovod/horovod)\n* Automated mixed precision [AMP](https://docs.nvidia.com/deeplearning/performance/mixed-precision-training/index.html)\n\nThis model is trained with mixed precision using Tensor Cores on Volta, Turing, and the NVIDIA Ampere GPU architectures. Therefore, researchers can get results 3x faster than training without Tensor Cores, while experiencing the benefits of mixed precision training. This model is tested against each NGC monthly container release to ensure consistent accuracy and performance over time.\n\n### Default configuration\n\nThe following sections highlight the default configuration for the ResNet50 model.\n\n#### Optimizer\n\nThis model uses the SGD optimizer with the following hyperparameters:\n\n* Momentum (0.875).\n* Learning rate (LR) = 0.256 for 256 batch size, for other batch sizes we linearly scale the learning\n rate. \n* Learning rate schedule - we use cosine LR schedule.\n* For bigger batch sizes (512 and up) we use linear warmup of the learning rate.\nduring the first 5 epochs according to [Training ImageNet in 1 hour](https://arxiv.org/abs/1706.02677).\n* Weight decay: 3.0517578125e-05 (1/32768).\n* We do not apply Weight decay on batch norm trainable parameters (gamma/bias).\n* Label Smoothing: 0.1.\n* We train for:\n * 90 Epochs -> 90 epochs is a standard for ResNet family networks.\n * 250 Epochs -> best possible accuracy. \n* For 250 epoch training we also use [MixUp regularization](https://arxiv.org/pdf/1710.09412.pdf).\n\n#### Data Augmentation\n\nThis model uses the following data augmentation:\n\n* For training:\n * Normalization.\n * Random resized crop to 224x224.\n * Scale from 8% to 100%.\n * Aspect ratio from 3/4 to 4/3.\n * Random horizontal flip.\n* For inference:\n * Normalization.\n * Scale to 256x256.\n * Center crop to 224x224.\n\n### Feature support matrix\n\nThe following features are supported by this model.\n\n| Feature | ResNet-50 v1.5 Tensorflow |\n|-----------------------|--------------------------\n|Multi-GPU training with [Horovod](https://github.com/horovod/horovod) | Yes |\n|[NVIDIA DALI](https://docs.nvidia.com/deeplearning/dali/release-notes/index.html) | Yes |\n|Automatic mixed precision (AMP) | Yes |\n\n#### Features\n\nMulti-GPU training with Horovod - Our model uses Horovod to implement efficient multi-GPU training with NCCL.\nFor details, refer to the example sources in this repository or the [TensorFlow tutorial](https://github.com/horovod/horovod/#usage).\n\nNVIDIA DALI - DALI is a library accelerating data preparation pipeline. To accelerate your input pipeline, you only need to define your data loader\nwith the DALI library. For details, refer to the example sources in this repository or the [DALI documentation](https://docs.nvidia.com/deeplearning/dali/index.html).\n\nAutomatic mixed precision (AMP) - Computation graph can be modified by TensorFlow on runtime to support mixed precision training. \nDetailed explanation of mixed precision can be found in the next section.\n\n### Mixed precision training\nMixed precision is the combined use of different numerical precisions in a computational method. [Mixed precision](https://arxiv.org/abs/1710.03740) training offers significant computational speedup by performing operations in half-precision format while storing minimal information in single-precision to retain as much information as possible in critical parts of the network. Since the introduction of [Tensor Cores](https://developer.nvidia.com/tensor-cores) in Volta, and following with both the Turing and Ampere architectures, significant training speedups are experienced by switching to mixed precision -- up to 3x overall speedup on the most arithmetically intense model architectures. Using [mixed precision training](https://docs.nvidia.com/deeplearning/performance/mixed-precision-training/index.html) previously required two steps:\n\n1. Porting the model to use the FP16 data type where appropriate. \n2. Adding loss scaling to preserve small gradient values.\n\nThis can now be achieved using Automatic Mixed Precision (AMP) for TensorFlow to enable the full [mixed precision methodology](https://docs.nvidia.com/deeplearning/sdk/mixed-precision-training/index.html#tensorflow) in your existing TensorFlow model code. AMP enables mixed precision training on Volta and Turing GPUs automatically. The TensorFlow framework code makes all necessary model changes internally.\n\nIn TF-AMP, the computational graph is optimized to use as few casts as necessary and maximize the use of FP16, and the loss scaling is automatically applied inside of supported optimizers. AMP can be configured to work with the existing tf.contrib loss scaling manager by disabling the AMP scaling with a single environment variable to perform only the automatic mixed-precision optimization. It accomplishes this by automatically rewriting all computation graphs with the necessary operations to enable mixed precision training and automatic loss scaling.\n\nFor information about:\n- How to train using mixed precision, see the [Mixed Precision Training](https://arxiv.org/abs/1710.03740) paper and [Training With Mixed Precision](https://docs.nvidia.com/deeplearning/performance/mixed-precision-training/index.html) documentation.\n- Techniques used for mixed precision training, see the [Mixed-Precision Training of Deep Neural Networks](https://devblogs.nvidia.com/mixed-precision-training-deep-neural-networks/) blog.\n- How to access and enable AMP for TensorFlow, see [Using TF-AMP](https://docs.nvidia.com/deeplearning/dgx/tensorflow-user-guide/index.html#tfamp) from the TensorFlow User Guide.\n\n#### Enabling mixed precision\n\nMixed precision is enabled in TensorFlow by using the Automatic Mixed Precision (TF-AMP) extension which casts variables to half-precision upon retrieval, while storing variables in single-precision format. Furthermore, to preserve small gradient magnitudes in backpropagation, a [loss scaling](https://docs.nvidia.com/deeplearning/sdk/mixed-precision-training/index.html#lossscaling) step must be included when applying gradients. In TensorFlow, loss scaling can be applied statically by using simple multiplication of loss by a constant value or automatically, by TF-AMP. Automatic mixed precision makes all the adjustments internally in TensorFlow, providing two benefits over manual operations. First, programmers need not modify network model code, reducing development and maintenance effort. Second, using AMP maintains forward and backward compatibility with all the APIs for defining and running TensorFlow models.\n\nTo enable mixed precision, you can simply add the values to the environmental variables inside your training script:\n- Enable TF-AMP graph rewrite:\n ```\n os.environ[\"TF_ENABLE_AUTO_MIXED_PRECISION_GRAPH_REWRITE\"] = \"1\"\n ```\n \n- Enable Automated Mixed Precision:\n ```\n os.environ['TF_ENABLE_AUTO_MIXED_PRECISION'] = '1'\n\n ```\n\n#### Enabling TF32\n\nTensorFloat-32 (TF32) is the new math mode in [NVIDIA A100](https://www.nvidia.com/en-us/data-center/a100/) GPUs for handling the matrix math also called tensor operations. TF32 running on Tensor Cores in A100 GPUs can provide up to 10x speedups compared to single-precision floating-point math (FP32) on Volta GPUs. \n\nTF32 Tensor Cores can speed up networks using FP32, typically with no loss of accuracy. It is more robust than FP16 for models which require high dynamic range for weights or activations.\n\nFor more information, refer to the [TensorFloat-32 in the A100 GPU Accelerates AI Training, HPC up to 20x](https://blogs.nvidia.com/blog/2020/05/14/tensorfloat-32-precision-format/) blog post.\n\nTF32 is supported in the NVIDIA Ampere GPU architecture and is enabled by default.","displayName","ResNet50 v1.5 for TensorFlow1",["D",1667525014116],"publisher","NVIDIA Deep Learning Examples","latestVersion","collectionCounts","imageUrl","/api/image/nvidia/dle/resources/resnet_tf1","rawImageUrl","https://assets.nvidiagrid.net/ngc/logos/Deep-Learning-Examples.png","canGuestDownload","latestVersionSizeInBytes","licenseTerms","isMultinodeEnabled","isMultiArchitecture","bias","explainability","privacy","safetyAndSecurity","featureLabels",[],"displayLabels",[],"policyLabels",[],"systemLabels",[],"products",[],"hasNimMetadata",{"_151":1256,"_153":1257},"ngcApps",[1209],"csaApiBaseUrl","https://www.nvidia.com/ai-agents/csa-agent/dcsm","datadogRum",{"_1190":1191,"_1192":1193,"_1194":1195,"_1196":1197,"_1198":285,"_109":311,"_1199":1200,"_1201":59,"_1202":90,"_1203":90,"_1204":1205,"_1206":90,"_1207":1208},"amplitude",{"_1187":1188,"_824":-7,"_1189":34},"disableJs","environment","prod","featureFlagsBlob","JTdCJTIydGVzdEZsYWclMjIlM0FmYWxzZSUyQyUyMnNob3dTZWFyY2hMYWJlbFNldEZpbHRlcnMlMjIlM0FmYWxzZSUyQyUyMmVuYWJsZUNzYVdpZGdldCUyMiUzQWZhbHNlJTJDJTIyZW5hYmxlUHJvZHVjdEdyb3VwaW5nQnlTdWJzY3JpcHRpb24lMjIlM0FmYWxzZSUyQyUyMmNvbnRhaW5lclNjYW5uaW5nUmVkZXNpZ25NMiUyMiUzQWZhbHNlJTdE","featureFlagsGuarded","isDevLikeEnv","isPlaywright","isMobileDevice","categorizedOrgs",{"_1181":1182,"_1183":1184,"_1185":1186},"currentOrgContext","currentContextRoles",[],"enterpriseOrgs",[],"manageableExternalUserGroups",[],"ngcApiOrigin","ngcHelmOrigin","origin","https://catalog.ngc.nvidia.com","usesMocks","theme","year",2026,"now",1789368739099,"1.15.0","defaultSearchSuggestions",{"_1064":391,"_1065":1066,"_1067":1068,"_1069":1070},"announcements",[],"consentState",{"_348":1054,"_1055":90,"_1056":34,"_1057":34,"_1058":1059},"consentConfig",{"_330":331,"_332":333,"_334":335,"_336":337,"_338":339,"_340":341,"_342":343,"_344":345,"_346":347,"_348":349},"showContactAdmin",1789368739100,[],[],[325,326,327,328,329],{"_48":63,"_50":73,"_52":146,"_54":-5,"_56":-5,"_58":59},{"_48":63,"_50":70,"_52":146,"_54":-5,"_56":-5,"_58":59},{"_48":63,"_50":64,"_52":146,"_54":-5,"_56":-5,"_58":59},{"_48":60,"_50":61,"_52":146,"_54":-5,"_56":-5,"_58":59},{"_48":49,"_50":51,"_52":146,"_54":-5,"_56":-5,"_58":59},"mainInfoText","We and our third-party partners (including social media, advertising, and analytics partners) use cookies and other tracking technologies to collect, store, monitor, and process certain information about you when you visit our website. The information collected might relate to you, your preferences, or your device. We use that information to make the site work, analyze performance and traffic on our website, provide a more personalized web experience, and assist in our marketing efforts.

Under certain privacy laws, you have the right to direct us not to \"sell\" or \"share\" your personal information for targeted advertising. To opt-out of the \"sale\" and \"sharing\" of personal information through cookies, you must opt-out of optional cookies using the toggles below. To opt out of the \"sale\" and \"sharing\" of data collected by other means (e.g., online forms) you must also update your data sharing preferences through the NVIDIA Preference Center.

Click on the different category headings below to find out more and change the settings according to your preference. You cannot opt out of Required Cookies as they are deployed to ensure the proper functioning of our website (such as prompting the cookie banner and remembering your settings, etc.). By clicking \"Save and Accept\" or \"Decline All\" at the bottom, you consent to the use of cookies and other tools as described in our Cookie Policy in accordance with your settings and accept our Terms of Service (which contains important waivers). For more information about our privacy practices, please see our Privacy Policy.","mainText","Cookie Settings","alertNoticeText",{"_1046":1047,"_1048":1049,"_1050":1051,"_1052":1053},"bannerAcceptLabel","Accept All","bannerRejectLabel","Reject Optional","bannerSettingsLabel","Manage Settings","preferenceSaveLabel","Save Choices","preferenceRejectLabel","Decline All","alwaysActiveText","Always Active","groups",[350,351,352,353,354],{"_355":736,"_112":737,"_211":738,"_359":90,"_360":16,"_362":739,"_364":740},{"_355":638,"_112":639,"_211":640,"_359":34,"_360":641,"_362":642,"_364":643},{"_355":546,"_112":547,"_211":548,"_359":34,"_360":62,"_362":549,"_364":550},{"_355":366,"_112":367,"_211":368,"_359":34,"_360":369,"_362":370,"_364":371},{"_355":356,"_112":357,"_211":358,"_359":34,"_360":361,"_362":363,"_364":365},"optanonGroupId","C0005","Social Media Cookies","These cookies are set by a range of social media services that we have added to the site to enable you to share our content with your friends and networks. They are capable of tracking your browser across other sites and building up a profile of your interests. This may impact the content and messages you see on other websites you visit. If you do not allow these cookies you may not be able to use or see these sharing tools.","alwaysActive","order",11,"cookies",[],"hosts",[],"C0004","Advertising Cookies","These cookies record your visit to our websites, the pages you have visited and the links you have followed to influence the advertisements that you see on other websites. These cookies and the information they collect may be managed by other companies, including our advertising partners, and may be used to build a profile of your interests and show you relevant advertising on other sites. We and our advertising partners will use this information to make our websites and the advertising displayed on it, more relevant to your interests.",5,[488,489,490,491,492,493,494,495,496,497,498,499,500,501,502,503,504,505,506,507,508,509],[372,373,374,375,376,377,378,379,380,381,382,383,384,385,386,387,388],{"_389":483,"_243":483,"_211":391,"_362":484},{"_389":479,"_243":479,"_211":391,"_362":480},{"_389":474,"_243":474,"_211":391,"_362":475},{"_389":470,"_243":470,"_211":391,"_362":471},{"_389":465,"_243":465,"_211":391,"_362":466},{"_389":460,"_243":460,"_211":391,"_362":461},{"_389":452,"_243":452,"_211":391,"_362":453},{"_389":444,"_243":444,"_211":391,"_362":445},{"_389":439,"_243":439,"_211":391,"_362":440},{"_389":434,"_243":434,"_211":391,"_362":435},{"_389":429,"_243":429,"_211":391,"_362":430},{"_389":424,"_243":424,"_211":391,"_362":425},{"_389":419,"_243":419,"_211":391,"_362":420},{"_389":412,"_243":412,"_211":391,"_362":413},{"_389":407,"_243":407,"_211":391,"_362":408},{"_389":400,"_243":400,"_211":391,"_362":401},{"_389":390,"_243":390,"_211":391,"_362":392},"hostName","amazon-adsystem.com","",[393,394],{"_112":399,"_396":390,"_397":34,"_211":398},{"_112":395,"_396":390,"_397":34,"_211":398},"ad-privacy","host","isSession","This domain is owned by online retailer Amazon and is used as part of its affiliate marketing programme.","ad-id","doubleclick.net",[402,403],{"_112":406,"_396":400,"_397":34,"_211":405},{"_112":404,"_396":400,"_397":34,"_211":405},"IDE","This domain is owned by Doubleclick (Google). The main business activity is: Doubleclick is Googles real time bidding advertising exchange","test_cookie","bing.com",[409],{"_112":410,"_396":407,"_397":34,"_211":411},"MUID","This domain is owned by Mircosoft - it is the site for the search engine Bing.","assets.adobedtm.com",[414,415],{"_112":418,"_396":412,"_397":34,"_211":417},{"_112":416,"_396":412,"_397":34,"_211":417},"kndctr_F207D74D549850760A4C98C6_AdobeOrg_*","If you know anything about this host and its uses please add your contributions here. ","kndctr_F207D74D549850760A4C98C6_AdobeOrg_consent","tiktok.com",[421],{"_112":422,"_396":419,"_397":34,"_211":423},"_ttp","TikTok is a social media platform owned by ByteDance, known for short-form video sharing. The service is used to create and share videos, and it utilizes cookies for various purposes including advertising and analytics.","t.co",[426],{"_112":427,"_396":424,"_397":34,"_211":428},"muc_ads","This domain is owned by Twitter. It is used as a URL shortening service, which allows Twitter to collect data on how links are shared.","snowplow.com",[431],{"_112":432,"_396":429,"_397":34,"_211":433},"sp","Testing","google.com",[436],{"_112":437,"_396":434,"_397":34,"_211":438},"NID","This domain is owned by Google Inc. Although Google is primarily known as a search engine, the company provides a diverse range of products and services. Its main source of revenue however is advertising. Google tracks users extensively both through its own products and sites, and the numerous technologies embedded into many millions of websites around the world. It uses the data gathered from most of these services to profile the interests of web users and sell advertising space to organisations based on such interest profiles as well as aligning adverts to the content on the pages where its customer's adverts appear.","linkedin.com",[441],{"_112":442,"_396":439,"_397":34,"_211":443},"bcookie","This domain is owned by LinkedIn, the business networking platform. It typically acts as a third party host where website owners have placed one of its content sharing buttons in their pages, although its content and services can be embedded in other ways. Although such buttons add functionality to the website they are on, cookies are set regardless of whether or not the visitor has an active Linkedin profile, or agreed to their terms and conditions. For this reason it is classified as a primarily tracking/targeting domain.","twitter.com",[446,447],{"_112":450,"_396":444,"_397":34,"_211":451},{"_112":448,"_396":444,"_397":34,"_211":449},"personalization_id","This domain is owned by Twitter. The main business activity is: Social Networking Services. Where twitter acts as a third party host, it collects data through a range of plug-ins and integrations, that is primarily used for tracking and targeting.","guest_*","This domain is owned by X (formerly Twitter). The main business activity is: Social Networking Services. Where X (formerly Twitter) acts as a third party host, it collects data through a range of plug-ins and integrations, that is primarily used for tracking and targeting.","youtube.com",[454,455],{"_112":458,"_396":452,"_397":34,"_211":459},{"_112":456,"_396":452,"_397":34,"_211":457},"VISITOR_INFO1_LIVE","This cookie is used as a unique identifier to track viewing of videos","__Secure-*","YouTube is a Google owned platform for hosting and sharing videos. YouTube collects user data through videos embedded in websites, which is aggregated with profile data from other Google services in order to display targeted advertising to web visitors across a broad range of their own and other websites.","demdex.net",[462],{"_112":463,"_396":460,"_397":34,"_211":464},"demdex","This cookie helps Adobe Audience Manger perform basic functions such as visitor identification, ID synchronization, segmentation, modeling, reporting, etc.","im-apps.net",[467],{"_112":468,"_396":465,"_397":34,"_211":469},"receive-cookie-deprecation","This domain is owned by im-apps.net which provides information and resources about various instant messaging (IM) applications to help users make informed decisions based on their communication needs.","influ2.com",[472],{"_112":473,"_396":470,"_397":34,"_211":433},"R","6sc.co",[476],{"_112":477,"_396":474,"_397":34,"_211":478},"6suuid","This domain is associated with the company 6Sense Insights, Inc., which provides predictive analytics and account engagement services. It is primarily used for tracking and targeting in marketing automation and advertising.","hushly.com",[481],{"_112":482,"_396":479,"_397":34,"_211":433},"_hly_gvid","t.contentsquare.net",[485],{"_112":486,"_396":483,"_397":90,"_211":487},"_cs_*","This is a session cookie. This is a pattern type cookie where a random 13-digit number is appended to the prefix _cs_.",{"_112":544,"_396":511,"_397":34,"_211":545},{"_112":542,"_396":511,"_397":34,"_211":543},{"_112":540,"_396":511,"_397":34,"_211":541},{"_112":422,"_396":511,"_397":34,"_211":539},{"_112":516,"_396":511,"_397":90,"_211":518},{"_112":538,"_396":511,"_397":34,"_211":391},{"_112":537,"_396":514,"_397":34,"_211":391},{"_112":535,"_396":511,"_397":34,"_211":536},{"_112":533,"_396":511,"_397":34,"_211":534},{"_112":532,"_396":511,"_397":34,"_211":391},{"_112":530,"_396":511,"_397":34,"_211":531},{"_112":516,"_396":529,"_397":90,"_211":518},{"_112":527,"_396":514,"_397":34,"_211":528},{"_112":527,"_396":517,"_397":34,"_211":528},{"_112":486,"_396":511,"_397":34,"_211":487},{"_112":525,"_396":511,"_397":34,"_211":526},{"_112":432,"_396":523,"_397":34,"_211":524},{"_112":521,"_396":511,"_397":34,"_211":522},{"_112":519,"_396":511,"_397":34,"_211":520},{"_112":516,"_396":517,"_397":90,"_211":518},{"_112":513,"_396":514,"_397":34,"_211":515},{"_112":510,"_396":511,"_397":90,"_211":512},"AMCVS_xxxxxAdobeOrg","nvidia.com","This is a pattern type cookie name associated with Adobe Marketing Cloud. It stores a unique visitor identifier, and uses an organisation identifier to allow a company to track users across their domains and services.This is a pattern type cookie name associated with Adobe Marketing Cloud. It stores a unique visitor identifier, and uses an organisation identifier.","_rdt_pn","www.nvidia.com","This cookie name is associated with www.mind.org.uk which is used for marketing purposes.","AMP_*","build.nvidia.com","Supports Amplitude product analytics.","_gcl_au","Used by Google AdSense for experimenting with advertisement efficiency across websites using their services","_mkto_trk","This cookie is associated with an email marketing service provided by Marketo. This tracking cookie allows a website to link visitor behaviour to the recipient of an email marketing campaign, to measure campaign effectiveness.","investor.nvidia.com","This cookie is associated with Eventbrite and is used to deliver content tailored to the end user's interests and improve content creation. This cookie is also used for event-booking purposes.","_gat_gtag_UA_140825383_2","Google Analytics","_rdt_uuid","This cookie is set by Reddit and is used for tracking the views of embedded videos and performance of the advertisements.","catalog.ngc.nvidia.com","_uetsid","This cookie is used by Bing to determine what ads should be shown that may be relevant to the end user perusing the site. ","kndctr_F207D74D549850760A4C98C6_AdobeOrg_bc_session_id","ttcsid_XXXXXXXXXXXXXXXXXXXXCategory","This cookie name is associated with TikTok which is used to track your activity on the platform, including how you interact with their embedded services on other websites.This cookie is associated with TikTok. It helps track your activity on the platform, including how you interact with their embedded services on other websites.","_fbp","Used by Facebook to deliver a series of advertisement products such as real time bidding from third party advertisers","_zitok","_twpid","This cookie is typically associated with the use of Taptap Digital, a company specializing in digital advertising. It is used for tracking and ad personalization purposes.","_tt_enable_cookie","This cookie name is associated with TikTok which is used to collect data about behavior and purchases on the website and to measure the effect of advertising. It is also used to track website activity to help optimize advertising efforts.","AMCV_xxxxxAdobeOrg","This cookie is used to identify a unique visitor for tracking purposes by Adobe Experience Cloud.This is a pattern type cookie name associated with Adobe Marketing Cloud. It stores a unique visitor identifier, and uses an organisation identifier to allow a company to track users across their domains and services.","ttcsid","This cookie name is associated with TikTok which is used to enable TikTok Ads session tracking by assigning a unique session ID to users who click on TikTok ads, allowing for the attribution of user behavior and conversions within the same session.This cookie is associated with TikTok. It helps track your activity on the platform, including how you interact with their embedded services on other websites.","C0003","Personalization Cookies","These cookies collect data about how you have interacted with our website to help us improve your web experience, such as which pages you have visited. These cookies may store a unique ID so that our system will remember you when you return. They may be set by us or by third party providers whose services we have added to our pages. These cookies enable us to provide enhanced website functionality and personalization as well as make the marketing messages we send to you more relevant to your interests. If you do not allow these cookies, then some or all of these services may not function properly.",[580,581,582,583,584,585,586,587,588,589,590,591,592,593,594,595,596,597,598,599,600,601,602,603,604],[551,552,553,554,555],{"_389":465,"_243":465,"_211":391,"_362":575},{"_389":439,"_243":439,"_211":391,"_362":570},{"_389":565,"_243":565,"_211":391,"_362":566},{"_389":560,"_243":560,"_211":391,"_362":561},{"_389":556,"_243":556,"_211":391,"_362":557},"nvidia.partners",[558],{"_112":559,"_396":556,"_397":34,"_211":433},"edge","app.hushly.com",[562],{"_112":563,"_396":560,"_397":34,"_211":564},"*CORS","This domain is associated with Hushly, a platform that provides lead conversion solutions for content marketing by using personalization and content engagement strategies.","github.com",[567],{"_112":568,"_396":565,"_397":34,"_211":569},"_octo","This domain is owned by Github. The main business activity is: Source code repository and collaboration platform",[571,572],{"_112":574,"_396":439,"_397":34,"_211":443},{"_112":573,"_396":439,"_397":34,"_211":443},"lidc","li_gc",[576,577],{"_112":579,"_396":465,"_397":34,"_211":469},{"_112":578,"_396":465,"_397":34,"_211":469},"imid_*","imid",{"_112":605,"_396":514,"_397":34,"_211":607},{"_112":636,"_396":511,"_397":34,"_211":637},{"_112":605,"_396":635,"_397":34,"_211":607},{"_112":633,"_396":511,"_397":90,"_211":634},{"_112":631,"_396":511,"_397":34,"_211":632},{"_112":630,"_396":613,"_397":90,"_211":391},{"_112":605,"_396":609,"_397":34,"_211":607},{"_112":628,"_396":511,"_397":34,"_211":629},{"_112":605,"_396":517,"_397":34,"_211":607},{"_112":616,"_396":514,"_397":34,"_211":391},{"_112":605,"_396":627,"_397":34,"_211":607},{"_112":625,"_396":511,"_397":34,"_211":626},{"_112":622,"_396":623,"_397":34,"_211":624},{"_112":620,"_396":621,"_397":34,"_211":391},{"_112":605,"_396":617,"_397":34,"_211":607},{"_112":619,"_396":511,"_397":34,"_211":391},{"_112":605,"_396":618,"_397":34,"_211":607},{"_112":616,"_396":617,"_397":34,"_211":391},{"_112":615,"_396":613,"_397":90,"_211":391},{"_112":605,"_396":614,"_397":34,"_211":607},{"_112":605,"_396":613,"_397":34,"_211":607},{"_112":605,"_396":612,"_397":34,"_211":607},{"_112":611,"_396":511,"_397":90,"_211":607},{"_112":608,"_396":609,"_397":90,"_211":610},{"_112":605,"_396":606,"_397":34,"_211":607},"_gd_*","info.nvidia.com","This cookie contains no identifying information and is used for diagnostic purposes by Impact Radius, a service we use to track signup commissions for our affiliate program. We have categorized it under Functionality because it is necessary for the operation of our affiliate program.","nv_country_code","developer.nvidia.com","This cookie is likely used to store the country code of the visitor to provide localized content and possibly for legal compliance with regional data privacy laws.","_gdxxxxxxxxxxxxx","nvidianews.nvidia.com","docs.api.nvidia.com","research.nvidia.com","readme_library","wcmmode","ai-podcast.nvidia.com","resources.nvidia.com","geo_locator","openedx-language-preference","learn.nvidia.com","doxygen_width","raytracing-docs.nvidia.com","This cookie is likely used to remember the width preferences of a sidebar or interface element within a documentation platform like Doxygen.","_sfid_eb78","This cookie is categorized as Functionality Cookie.","crowdsource.nvidia.com","mbox","Cookie set by Adobe SiteCatalyst analytics software - used for measuring the performance of page content using A/B split testing","readme_language","vid","This cookie name 'vid' is used for tracking the search request made by the user on the 'Find a practitioner' page.","at_check","Used by Adobe Target to check if cookies are enabled/supported on the browser","blogs.nvidia.com","nvHeadless","This cookie is likely used to track if a user is using the site in a 'headless' non-GUI mode, which could be for accessibility reasons or for automated processes.","C0002","Performance Cookies","These cookies are used to provide quantitative measures of our website visitors, such as the number of times you visit, time on page, your mouse movements, scrolling, clicks and keystroke activity on the websites; other browsing, search, or product research behavior; and what brought you to our site. These cookies may store a unique ID so that our system will remember you when you return. Information collected with these cookies is used to measure and find ways to improve website performance.",3,[664,665,666,667,668,669,670,671,672,673,674,675,676,677,678,679,680,681,682,683,684,685,686,687,688,689,690,691,692],[644,645,646,647],{"_389":659,"_243":659,"_211":391,"_362":660},{"_389":654,"_243":654,"_211":391,"_362":655},{"_389":452,"_243":452,"_211":391,"_362":651},{"_389":565,"_243":565,"_211":391,"_362":648},[649],{"_112":650,"_396":565,"_397":90,"_211":569},"_gh_sess",[652],{"_112":653,"_396":452,"_397":90,"_211":459},"YSC","bizible.com",[656],{"_112":657,"_396":654,"_397":34,"_211":658},"_BUID","Bizible is a marketing attribution software that tracks the performance of marketing efforts and connects them to revenue, primarily used for B2B environments.","a.usbrowserspeed.com",[661],{"_112":662,"_396":659,"_397":34,"_211":663},"tuid","This domain is owned by Ookla which provides internet speed testing services through its Speedtest application.",{"_112":734,"_396":511,"_397":34,"_211":735},{"_112":719,"_396":733,"_397":90,"_211":720},{"_112":702,"_396":514,"_397":34,"_211":626},{"_112":702,"_396":614,"_397":34,"_211":626},{"_112":731,"_396":511,"_397":34,"_211":732},{"_112":719,"_396":517,"_397":90,"_211":720},{"_112":729,"_396":511,"_397":34,"_211":730},{"_112":719,"_396":514,"_397":90,"_211":720},{"_112":727,"_396":613,"_397":34,"_211":728},{"_112":719,"_396":635,"_397":90,"_211":720},{"_112":699,"_396":514,"_397":34,"_211":701},{"_112":725,"_396":511,"_397":90,"_211":726},{"_112":723,"_396":511,"_397":34,"_211":724},{"_112":721,"_396":511,"_397":90,"_211":722},{"_112":719,"_396":609,"_397":90,"_211":720},{"_112":717,"_396":511,"_397":34,"_211":718},{"_112":715,"_396":511,"_397":34,"_211":716},{"_112":713,"_396":511,"_397":90,"_211":714},{"_112":711,"_396":618,"_397":34,"_211":712},{"_112":699,"_396":609,"_397":34,"_211":701},{"_112":709,"_396":511,"_397":34,"_211":710},{"_112":707,"_396":511,"_397":90,"_211":708},{"_112":705,"_396":511,"_397":34,"_211":706},{"_112":703,"_396":511,"_397":34,"_211":704},{"_112":702,"_396":511,"_397":34,"_211":626},{"_112":699,"_396":700,"_397":34,"_211":701},{"_112":697,"_396":511,"_397":34,"_211":698},{"_112":695,"_396":511,"_397":90,"_211":696},{"_112":693,"_396":511,"_397":34,"_211":694},"_ga_xxxxxxxxxx","Used by Google Analytics to identify and track an individual session with your device.","s_ppvl","This cookie is set by Adobe Analytics to store information on the percentage of the page displayed to the end user.","coveo_visitorId","This cookie name is associated with Coveo which is used to store a unique user ID for marketing purposes and to track the number of distinct users who submitted a query or clicked an item, counting them only once even if they did the same query or clicked the same item over the course of many visits.","_dd_s","marketplace.nvidia.com","Datadog Cookie used to group all events generated from a unique user session across multiple pages. It contains the current session ID, whether the session is excluded due to sampling, and the expiration date of the session. The cookie is extended for an extra 15 minutes every time the user interacts with the website, up to the maximum user session duration (4 hours).","nvweb_A","_ga","This cookie name is associated with Google Universal Analytics - which is a significant update to Google's more commonly used analytics service. This cookie is used to distinguish unique users by assigning a randomly generated number as a client identifier. It is included in each page request in a site and used to calculate visitor, session and campaign data for the sites analytics reports. By default it is set to expire after 2 years, although this is customisable by website owners.","_uetvid","This is a cookie utilised by Microsoft Bing Ads and is a tracking cookie. It allows us to engage with a user that has previously visited our website.","tp","This cookie is associated with Adobe Analytics site tracking. It registers if the website's scroll-depth detection is active. This function remembers how far the user has scrolled on the current session across the website's sub-pages.","_hly_vid","This cookie _hly_vid is used to record website and server performance with each visit to the site. The data is aggregated and analyzed; therefore, it remains anonymous upon use.","_session_id","This cookie creates an interim session ID used as an in-session user ID.\n","s_cc","This cookie is associated with the Adobe Site Catalyst. It determines whether cookies are enabled in the web browser.","s_ecid","Adobe Analytics cookie containing the the Adobe Experience Cloud 1st party visitor id if the Analytics AMCV visitor id is not found.","_hjSessionUser_xxxxxx","Hotjar cookie that is set when a user first lands on a page with the Hotjar script. It is used to persist the Hotjar User ID, unique to that site on the browser. This ensures that behavior in subsequent visits to the same site will be attributed to the same user ID.","_hjHasCachedUserAttributes","This cookie name is associated with Hotjar which is used to know whether the data set in _hjUserAttributes Local Storage item is up to date or not.","s_ppv","This cookie is associated with Adobe Analytics site tracking and is used to measure a user's scroll activity to see how much of a page they view before moving to another page.","_gid","This cookie name is associated with Google Universal Analytics. This appears to be a new cookie and as of Spring 2017 no information is available from Google. It appears to store and update a unique value for each page visited.","_hjTLDTest","When the Hotjar script executes we try to determine the most generic cookie path we should use, instead of the page hostname. This is done so that cookies can be shared across subdomains (where applicable). To determine this, we try to store the _hjTLDTest cookie for different URL substring alternatives until it fails. After this check, the cookie is removed.","ekfls-2025-03-27","This cookie name is associated with readme.io which is used to help maintain data for the API reference support page.","s_getNewRepeat","This cookie is used to determine whether a visitor is a new visitor or a repeat visitor to the site, and is used for analytical purposes.","_hly_sid","This cookie is likely associated with website analytics, as the prefix '_hly_' suggests a connection to Holistics, a business intelligence and analytics platform.","docs.nvidia.com","_hjSession_xxxxxx","A cookie that holds the current session data. This ensues that subsequent requests within the session window will be attributed to the same Hotjar session.","C0001","Required Cookies","These cookies enable core functionality such as security, network management, and accessibility. These cookies are required for the site to function and cannot be turned off.",[881,882,883,884,885,886,887,888,889,890,891,892,893,894,895,896,897,898,899,900,901,902,903,904,905,906,907,908,909,910,911,912,913,914,915,916,917,918,919,920,921,922,923,924,925,926,927,928,929,930,931,932,933,934,935,936,937,938,939,940,941,942,943,944,945,946,947,948,949,950,951,952,953],[741,742,743,744,745,746,747,748,749,750,751,752,753,754,755,756,757,758,759,760,761,762,763,764,765],{"_389":878,"_243":878,"_211":391,"_362":879},{"_389":872,"_243":872,"_211":391,"_362":873},{"_389":869,"_243":869,"_211":391,"_362":870},{"_389":857,"_243":857,"_211":391,"_362":858},{"_389":854,"_243":854,"_211":391,"_362":855},{"_389":851,"_243":851,"_211":391,"_362":852},{"_389":846,"_243":846,"_211":391,"_362":847},{"_389":452,"_243":452,"_211":391,"_362":843},{"_389":840,"_243":840,"_211":391,"_362":841},{"_389":444,"_243":444,"_211":391,"_362":838},{"_389":439,"_243":439,"_211":391,"_362":836},{"_389":831,"_243":831,"_211":391,"_362":832},{"_389":826,"_243":826,"_211":391,"_362":827},{"_389":821,"_243":821,"_211":391,"_362":822},{"_389":818,"_243":818,"_211":391,"_362":819},{"_389":565,"_243":565,"_211":391,"_362":815},{"_389":424,"_243":424,"_211":391,"_362":812},{"_389":560,"_243":560,"_211":391,"_362":805},{"_389":556,"_243":556,"_211":391,"_362":799},{"_389":793,"_243":793,"_211":391,"_362":794},{"_389":788,"_243":788,"_211":391,"_362":789},{"_389":783,"_243":783,"_211":391,"_362":784},{"_389":779,"_243":779,"_211":391,"_362":780},{"_389":774,"_243":774,"_211":391,"_362":775},{"_389":766,"_243":766,"_211":391,"_362":767},"polyfill.io",[768,769],{"_112":772,"_396":766,"_397":34,"_211":773},{"_112":770,"_396":766,"_397":34,"_211":771},"cf_ob_info","This cookie is associated with sites using CloudFlare. It is used to improve page load times and to override any security restrictions based on the visitors IP address. It does not contain any user identification information.","cf_use_ob","This cookie is associated with sites using CloudFlare. It is used to improve page load times and to override any security restrictions based on the visitors IP address. It does not contain any user identification information. ","www.recaptcha.net",[776],{"_112":777,"_396":774,"_397":34,"_211":778},"_GRECAPTCHA","This domain is used by Google reCAPTCHA, which provides spam protection services to websites by detecting if a user is a human or a bot.","api2.hcaptcha.com",[781],{"_112":782,"_396":779,"_397":34,"_211":433},"__cflb","instagram.com",[785],{"_112":786,"_396":783,"_397":34,"_211":787},"csrftoken","Instagram is a social networking service owned by Facebook, Inc. It's designed for sharing photos and videos from a smartphone. The service also uses cookies for various purposes, including analytics, personalization, and advertising.","readme.io",[790],{"_112":791,"_396":788,"_397":34,"_211":792},"__cf_bm","The __cf_bm cookie is a cookie necessary to support Cloudflare Bot Management, currently in private beta. As part of our bot management service, this cookie helps manage incoming traffic that matches criteria associated with bots.","nvidiacorp.us-5.evergage.com",[795,796],{"_112":798,"_396":793,"_397":34,"_211":433},{"_112":797,"_396":793,"_397":34,"_211":433},"AWSALBTG","AWSALBTGCORS",[800,801],{"_112":803,"_396":556,"_397":34,"_211":804},{"_112":802,"_396":556,"_397":34,"_211":433},"ak_bmsc","bm_mi","This domain is likely associated with NVIDIA's partner program, which could involve marketing, sales, and support services for partners.",[806,807,808],{"_112":811,"_396":560,"_397":34,"_211":433},{"_112":810,"_396":560,"_397":90,"_211":433},{"_112":809,"_396":560,"_397":34,"_211":433},"AWSALBCORS","JSESSIONID","AWSALB",[813],{"_112":791,"_396":424,"_397":34,"_211":814},"This is a CloudFoundry cookie",[816],{"_112":817,"_396":565,"_397":34,"_211":569},"logged_in","hcaptcha.com",[820],{"_112":791,"_396":818,"_397":34,"_211":814},"tracking.intentsify.io",[823],{"_112":824,"_396":821,"_397":34,"_211":825},"userId","Used to track usage of the embedded video player.","static.reo.dev",[828],{"_112":829,"_396":826,"_397":34,"_211":830},"cookietest","Common cookie name could have a number of different origins. Where this is first party and a session cookie, its most likely to do with checking to see if the browser is set to block or allow cookies.","www.datadoghq-browser-agent.com",[833],{"_112":834,"_396":831,"_397":34,"_211":835},"dd_cookie_test_","dd_cookie_test",[837],{"_112":791,"_396":439,"_397":34,"_211":814},[839],{"_112":791,"_396":444,"_397":34,"_211":814},"supabase.co",[842],{"_112":791,"_396":840,"_397":34,"_211":814},[844],{"_112":845,"_396":452,"_397":34,"_211":459},"VISITOR_PRIVACY_METADATA","dosa-prd.azurewebsites.net",[848],{"_112":849,"_396":846,"_397":34,"_211":850},"access_token","This cookie is used to obtain the permissions for users of the portal to allow for API access in order to retrieve information required to perform portal functions.","w.hcaptcha.com",[853],{"_112":791,"_396":851,"_397":34,"_211":814},"cdn.readme.io",[856],{"_112":791,"_396":854,"_397":34,"_211":814},"queue-it.net",[859,860,861,862,863],{"_112":868,"_396":857,"_397":34,"_211":391},{"_112":867,"_396":857,"_397":34,"_211":391},{"_112":866,"_396":857,"_397":34,"_211":391},{"_112":865,"_396":857,"_397":34,"_211":391},{"_112":864,"_396":857,"_397":34,"_211":391},"Queue-it-{queueId}","Queue-it-{dmonvidia }________{eventId}","Queue-it-token","queueitsoftblock_{queueId}","Queue-it-{nvidia}________{eventId}","x.com",[871],{"_112":791,"_396":869,"_397":34,"_211":792},"api.hcaptcha.com",[874,875],{"_112":877,"_396":872,"_397":34,"_211":433},{"_112":782,"_396":872,"_397":34,"_211":876},"This domain is associated with hCaptcha, a service that provides CAPTCHA and reCAPTCHA alternatives for user verification and bot protection.","hmt_id","vimeo.com",[880],{"_112":791,"_396":878,"_397":34,"_211":814},{"_112":1014,"_396":511,"_397":34,"_211":1045},{"_112":1043,"_396":606,"_397":90,"_211":1044},{"_112":1036,"_396":1005,"_397":90,"_211":1037},{"_112":1042,"_396":700,"_397":90,"_211":391},{"_112":1040,"_396":511,"_397":34,"_211":1041},{"_112":798,"_396":1026,"_397":34,"_211":965},{"_112":973,"_396":1035,"_397":90,"_211":975},{"_112":797,"_396":964,"_397":34,"_211":965},{"_112":976,"_396":1035,"_397":34,"_211":977},{"_112":1038,"_396":511,"_397":90,"_211":1039},{"_112":1036,"_396":1012,"_397":90,"_211":1037},{"_112":970,"_396":1035,"_397":34,"_211":972},{"_112":1033,"_396":511,"_397":34,"_211":1034},{"_112":797,"_396":956,"_397":34,"_211":965},{"_112":961,"_396":1032,"_397":34,"_211":963},{"_112":809,"_396":962,"_397":34,"_211":978},{"_112":811,"_396":962,"_397":34,"_211":1029},{"_112":976,"_396":971,"_397":34,"_211":977},{"_112":1030,"_396":609,"_397":34,"_211":1031},{"_112":811,"_396":635,"_397":34,"_211":1029},{"_112":1027,"_396":511,"_397":34,"_211":1028},{"_112":803,"_396":635,"_397":34,"_211":1022},{"_112":999,"_396":614,"_397":90,"_211":391},{"_112":797,"_396":1026,"_397":34,"_211":965},{"_112":1024,"_396":964,"_397":90,"_211":1025},{"_112":988,"_396":511,"_397":34,"_211":979},{"_112":961,"_396":511,"_397":34,"_211":963},{"_112":1023,"_396":523,"_397":34,"_211":391},{"_112":803,"_396":511,"_397":34,"_211":1022},{"_112":811,"_396":956,"_397":34,"_211":1021},{"_112":1019,"_396":618,"_397":90,"_211":1020},{"_112":1017,"_396":511,"_397":90,"_211":1018},{"_112":1016,"_396":511,"_397":34,"_211":391},{"_112":1014,"_396":962,"_397":34,"_211":1015},{"_112":1013,"_396":523,"_397":34,"_211":391},{"_112":1004,"_396":1012,"_397":90,"_211":1006},{"_112":1010,"_396":614,"_397":34,"_211":1011},{"_112":1008,"_396":700,"_397":34,"_211":1009},{"_112":1007,"_396":517,"_397":34,"_211":391},{"_112":1004,"_396":1005,"_397":90,"_211":1006},{"_112":1002,"_396":971,"_397":90,"_211":1003},{"_112":1000,"_396":511,"_397":34,"_211":1001},{"_112":999,"_396":514,"_397":90,"_211":391},{"_112":970,"_396":974,"_397":34,"_211":972},{"_112":802,"_396":989,"_397":34,"_211":979},{"_112":997,"_396":514,"_397":34,"_211":998},{"_112":995,"_396":511,"_397":34,"_211":996},{"_112":993,"_396":511,"_397":34,"_211":994},{"_112":992,"_396":511,"_397":34,"_211":391},{"_112":834,"_396":514,"_397":34,"_211":991},{"_112":988,"_396":635,"_397":34,"_211":979},{"_112":791,"_396":606,"_397":34,"_211":990},{"_112":988,"_396":989,"_397":34,"_211":979},{"_112":987,"_396":523,"_397":34,"_211":391},{"_112":986,"_396":511,"_397":34,"_211":959},{"_112":984,"_396":971,"_397":34,"_211":985},{"_112":797,"_396":982,"_397":34,"_211":983},{"_112":981,"_396":523,"_397":34,"_211":391},{"_112":980,"_396":511,"_397":34,"_211":391},{"_112":973,"_396":971,"_397":90,"_211":975},{"_112":802,"_396":635,"_397":34,"_211":979},{"_112":809,"_396":635,"_397":34,"_211":978},{"_112":976,"_396":974,"_397":34,"_211":977},{"_112":973,"_396":974,"_397":90,"_211":975},{"_112":970,"_396":971,"_397":34,"_211":972},{"_112":968,"_396":964,"_397":90,"_211":969},{"_112":966,"_396":621,"_397":34,"_211":967},{"_112":798,"_396":964,"_397":34,"_211":965},{"_112":961,"_396":962,"_397":34,"_211":963},{"_112":791,"_396":613,"_397":34,"_211":960},{"_112":958,"_396":511,"_397":34,"_211":959},{"_112":809,"_396":956,"_397":34,"_211":957},{"_112":954,"_396":511,"_397":34,"_211":955},"LoginHints","Keeps users logged in across NVIDIA domains.","api.ngc.nvidia.com","This cookie is managed by AWS and is used for load balancing.","bm_so","This cookie name is used to load the site as quickly as possible and to recognize bots using Akamai.","The __cf_bm cookie is a cookie necessary to support Cloudflare Bot Management, currently in private beta. As part of our bot management service, this cookie helps manage incoming traffic that matches criteria associated with bots.\nThis is a CloudFoundry cookie","OptanonConsent","nvid.nvidia.com","This cookie is set by the cookie compliance solution from OneTrust. It stores information about the categories of cookies the site uses and whether visitors have given or withdrawn consent for the use of each category. This enables site owners to prevent cookies in each category from being set in the users browser, when consent is not given. The cookie has a normal lifespan of one year, so that returning visitors to the site will have their preferences remembered. It contains no information that can identify the site visitor.","events.nvidia.com","This cookie is associated with the Amazon Web Services Elastic Load Balancing functionality for routing client request on the server.Used to attribute commission to affiliates when you arrive at the website from an affiliate referral link. It is set when you click on one of our links and used to let the advertiser and us know the website from which you came.","sessionid","This is a very generic cookie name that may have different purposes on different sites, but generally it will be some kind of anonymous session identifier.","_csrf","This cookie is used to prevent Cross-Site Request Forgery (CSRF) attacks, ensuring that each request made to the server is from the user's browser and is intentional.","LSKey-c$CookieConsentPolicy","enterprise-support.nvidia.com","Used to apply end-user cookie consent preferences set by our client-side utility.Custom- Short Duration. Unique ID of an Enquiry- used to validate access to case, and load exisiting data.\n","renderCtx","academicgrants.nvidia.com","Used to store site parameters in the session for reuse across requests by a single client for functionality and performance reasons.","CookieConsentPolicy","Used to apply end-user cookie consent preferences set by our client-side utility.","This cookie is managed by AWS and is used for load balancing.Used to attribute commission to affiliates when you arrive at the website from an affiliate referral link. It is set when you click on one of our links and used to let the advertiser and us know the website from which you came.","This cookie is associated with Akamai and is used to differentiate between traffic from humans and bots.","nv_co","__uxq412__ses.88c1","learn.learn.nvidia.com","This cookie is associated with the Amazon Web Services Elastic Load Balancing functionality for routing client request on the server.","pctrk","Used to count page views by unauthenticated users against license usage.","bm_s","bpazaws52gukakzc__ctrl0_ctl36_uccaptcha","bm_sv","store.nvidia.com","The __cf_bm cookie is a cookie necessary to support Cloudflare Bot Management, currently in private beta. As part of our bot management service, this cookie helps manage incoming traffic that matches criteria associated with bots.\n","This cookie name is used by DataDog to determine if cookies are supported on the website.dd_cookie_test","AMP_TLDTEST_xxxxxxxx","SESSlsd","This cookie is likely to be a session identifier used to maintain user session state across page requests.","mboxEdgeCluster","This cookie name is associated with Adobe Target which is used to communicate the correct edge cluster to at.js so that Target profiles can stay in sync as users navigate across a site.","_rdt_em","This cookie name is associated with Reddit which is used to manage session state and user interactions.","megaNavCollapsed","_abck","The _abck cookie is associated with Akamai's Bot Manager, which is used to prevent fraud and abuse by safeguarding online interactions and detecting irregularities that could suggest malicious activity, such as bot-driven attacks or hacking attempts.","idccsrf","This cookie name is associated with Salesforce which is used to track CrossSiteRequestForgery validation for certain single sign-on (SSO) flows.","ApplicationGatewayAffinity","partners.nvidia.com","This cookie is used to maintain a user session by the server in a load balanced environment, to ensure the user has a consistent experience.","AMP_TEST_xxxxxxxx","bm_lso","This cookie name is associated with Akamai which is used to load the site as quickly as possible and to recognize bots.","SSESS###########","This cookie is likely to be a session cookie used to maintain an anonymized user session by the server in a secure environment, often seen in sites built with the Drupal content management system.","nvdeveloper.nvidia.com","__uxq412__id.88c1","OptanonAlertBoxClosed","This cookie is set by websites using certain versions of the cookie law compliance solution from OneTrust. It is set after visitors have seen a cookie information notice and in some cases only when they actively close the notice down. It enables the website not to show the message more than once to a user. The cookie has a normal lifespan of one year and contains no personal information.","QueueITAccepted-SDFrts345E-V3_{eventId}","TEST_AMCV_COOKIE_WRITE","This cookie is used to test whether cookies are enabled in the browser. It is a temporary cookie used to determine if the browser accepts cookies.","XSRF-TOKEN","This cookie is written to help with site security in preventing Cross-Site Request Forgery attacks.This cookie helps with site security in preventing Cross-Site Request Forgery attacks.","AWS ELB application load balancer","This cookie is associated with Akamai. It is used to analyse traffic to determine if it is automated traffic generated by bots or a human user.","bpazaws52gukakzc__ctrl0_ctl78_list_ctl09_systemcaptcha","PHPSESSID","PHP session cookie associated with embedded content from this domain.\n","base.manager.learn.nvidia.com","bm_ss","This cookie is used to help protect websites from automated bot traffic.","AWS ELB application load balancerUsed to attribute commission to affiliates when you arrive at the website from an affiliate referral link. It is set when you click on one of our links and used to let the advertiser and us know the website from which you came.","_devzone_session","This cookie is likely used to maintain a user session on the website by the server. It is common for session identifiers to be used to keep users logged in and track their interaction with the site during a session.","docs.ngc.nvidia.com","bm_sz","This cookie is used by Akamai to optimize site performance and security. It is typically associated with the delivery of web content and services.","vcalliance.nvidia.com","ASP.NET_SessionId","General purpose platform session cookie, used by sites written with Microsoft .NET based technologies. Usually used to maintain an anonymised user session by the server.","com.adobe.alloy.getTld","This cookie name is associated with the Adobe WebSDK which is used to determine the top-most domain that can accept cookies.","SESSls","This cookie is likely to be a session identifier, used to maintain a user's session state across page requests.","c_code","BIGipServerx","This cookie name is associated with the BIG-IP product suite from company F5. Usually associated with managing sessions on load balanced servers, to ensure user requests are routed consistently to the correct server. The common root is BIGipServer most commonly followed by a domain name, usually the one that it is hosted on, but not always.","This cookie is set by websites using certain versions of the cookie law compliance solution from OneTrust. It is set after visitors have seen a cookie information notice and in some cases only when they actively close the notice down. It enables the website not to show the message more than once to a user. The cookie has a one year lifespan and contains no personal information.","default","

NVIDIA uses cookies to improve your experience on our web site. We and our third-party partners also use cookies and other tools to collect and record information you provide as well as information about your interactions with our websites for performance improvement, analytics, and to assist in marketing efforts. By clicking \"Accept All\", you consent to our use of cookies and other tools as described in our Cookie Policy. You can manage your cookie settings by clicking on \"Manage Settings.\" By continuing to use this site or by clicking one of the buttons below, you agree to our Terms of Service (which contains important waivers). Please see our Privacy Policy for more information on our privacy practices.

","gpc","

We have detected the Global Privacy Control (GPC) signal and recorded your rejection of all optional cookies on this site for this browser. You can manage your cookie settings by clicking on \"Manage Settings\". Please see our Cookie Policy for more information. To opt out of non-cookie personal information \"sales\" / \"sharing\" for targeted advertising purposes, please visit the NVIDIA Preference Center. Please see our Privacy Policy for more information on our privacy practices.

","gpcLoggedIn","

We have detected the Global Privacy Control Signal (GPC) and have opted you out of all optional cookies on this browser. You can manage your cookie settings by clicking on \"Manage Settings\". Please see our Cookie Policy for more information. We have also opted you out of \"sharing\"/\"sales\" of personal information outside of cookies. You can manage these settings in the NVIDIA NVIDIA Preference Center. Please see our Privacy Policy for more information.

","gpcLoggedInOverridden","

We have detected the Global Privacy Control Signal (GPC) and have opted you out of all optional cookies on this browser. You can manage your cookie settings by clicking on \"Manage Settings\". Please see our Cookie Policy for more information. We have also opted you out of \"sharing\"/\"sales\" of personal information outside of cookies which overrides at least one of your previous settings. You can manage them in the NVIDIA Preference Center. Please see our Privacy Policy for more information.

",{"_736":16,"_638":59,"_546":59,"_366":59,"_356":59},"hasConsented","gpcApplied","showBanner","geolocation",{"_1060":1061,"_1062":1063},"country","US","state","OR","query","categories",[1083,1084,1085,1086,1087],"optionUrls",[1071,1072,1073,1074,1075,1076,1077,1078,1079,1080,1081,1082],"total",12,"/orgs/nim/deepseek-ai/containers/deepseek-r1/-","/orgs/nim/nvidia/containers/llama-3.1-nemotron-70b-instruct/-","/orgs/nvidia/-/containers/pytorch/-","/orgs/nvidia/omniverse/collections/kit/-","/orgs/nvidia/-/collections/deepstream_sdk/-","/orgs/nvidia/omniverse/collections/kit-appstreaming-collection/-","/orgs/nvidia/research/models/stylegan3/-","/orgs/nvidia/tao/models/peoplenet/-","/orgs/nvidia/tao/models/trafficcamnet/-","/orgs/nvidia/riva/resources/riva_quickstart/-","/orgs/nvidia/-/helm-charts/gpu-operator/-","/orgs/nvidia/-/helm-charts/k8s-nim-operator/-",{"_20":1161,"_1089":1162,"_1091":59},{"_20":1126,"_1089":1127,"_1091":641},{"_20":1109,"_1089":1110,"_1091":1111},{"_20":21,"_1089":1102,"_1091":1103},{"_20":1088,"_1089":1090,"_1091":19},"helm-charts","items",[1092,1093],"startIndex",{"_20":1088,"_243":1099,"_205":206,"_207":-7,"_238":1100,"_246":1096,"_271":1101,"_1098":1081},{"_20":1088,"_243":1094,"_205":206,"_207":-7,"_238":1095,"_246":1096,"_271":1097,"_1098":1082},"NVIDIA NIM Operator","k8s-nim-operator","NVIDIA",[],"url","GPU Operator","gpu-operator",[],[1104],9,{"_20":21,"_243":1105,"_205":206,"_207":1106,"_238":1107,"_246":1096,"_271":1108,"_1098":1080},"Riva Skills Quick Start","riva","riva_quickstart",[],"models",[1112,1113,1114],6,{"_20":1109,"_243":1122,"_205":206,"_207":1123,"_238":1124,"_246":1096,"_271":1125,"_1098":1077},{"_20":1109,"_243":1119,"_205":206,"_207":1116,"_238":1120,"_246":1096,"_271":1121,"_1098":1078},{"_20":1109,"_243":1115,"_205":206,"_207":1116,"_238":1117,"_246":1096,"_271":1118,"_1098":1079},"TrafficCamNet","tao","trafficcamnet",[],"PeopleNet","peoplenet",[],"StyleGAN3 pretrained models","research","stylegan3",[],"collections",[1128,1129,1130],{"_20":1126,"_243":1158,"_205":206,"_207":1132,"_238":1159,"_246":1096,"_271":1160,"_1098":1074},{"_20":1126,"_243":1155,"_205":206,"_207":-7,"_238":1156,"_246":1096,"_271":1157,"_1098":1075},{"_20":1126,"_243":1131,"_205":206,"_207":1132,"_238":1133,"_246":1096,"_271":1134,"_1098":1076},"Omniverse Kit App Streaming","omniverse","kit-appstreaming-collection",[1135,1136,1137],{"_112":1149,"_243":1150,"_240":1151,"_250":1152,"_1142":1153},{"_112":1132,"_243":1145,"_240":1146,"_250":1141,"_1142":1147},{"_112":1138,"_243":1139,"_240":1140,"_250":1141,"_1142":1143},"omniverse-dev","NVIDIA Developer Program","Join the Developer Program for access to free tools, support, and tech resources.","https://store-images.s-microsoft.com/image/apps.33936.6f1bcf42-55d0-45d1-b4b6-f08c68a50bbe.e91ca707-db64-49d6-9761-4781590a9685.7ba64fc5-7c44-4864-a5e6-7bde05126935","supportedFeatures",[1144],"UNIFIED_CATALOG","NVIDIA Omniverse","NVIDIA Omniverse is a platform of APIs, services, and software development kits (SDKs) that enable developers to build generative AI-enabled tools and applications.",[1144,1148],"SHAREABLE_TO_USER_GROUPS","nv-ai-enterprise","NVIDIA AI Enterprise","Accelerate your AI agent development","https://assets.ngc.nvidia.com/products/catalog/images/product-NVAIE.jpg",[1144,1148,1154],"NGC_CONSOLE","DeepStream SDK","deepstream_sdk",[],"Omniverse Kit (FB)","kit",[1135,1136,1137],"containers",[1163,1164,1165],{"_20":1161,"_243":1177,"_205":1170,"_207":1178,"_238":1179,"_246":1096,"_271":1180,"_1098":1071},{"_20":1161,"_243":1169,"_205":1170,"_207":206,"_238":1171,"_246":1096,"_271":1172,"_1098":1072},{"_20":1161,"_243":1166,"_205":206,"_207":-7,"_238":1167,"_246":1096,"_271":1168,"_1098":1073},"PyTorch","pytorch",[],"Llama-3.1-Nemotron-70B-Instruct","nim","llama-3.1-nemotron-70b-instruct",[1135,1173],{"_112":1174,"_243":1139,"_240":1140,"_250":1175,"_1142":1176},"nim-dev","https://assets.ngc.nvidia.com/products/catalog/images/product-nim-dev.jpg",[1144],"DeepSeek-R1","deepseek-ai","deepseek-r1",[1135,1173],"ownerOrgs",[],"memberOrgs",[],"externalOrgs",[],"apiKey","2b44f86fea50e90ffd2db29ee2028419","consentGranted","applicationId","4c857e4f-5ffa-45fa-b75e-af773c22f0ad","clientToken","pubb6927db33dae60ba5031a186cde0242a","site","datadoghq.com","service","catalog-ui","env","sessionSampleRate",100,"sessionReplaySampleRate","compressIntakeRequests","storeContextsAcrossPages","defaultPrivacyLevel","mask-user-input","trackViewsManually","trackingConsent","not-granted",{"_1210":1211,"_1212":304,"_1213":34,"_1214":1215,"_1089":1216,"_1217":1218,"_1219":1211,"_1220":34,"_1221":34,"_1222":34,"_1223":1224},"appName","ngc-catalog","appOrigin","hideOnPrivateOrg","icon","ObjectsModules2",[1225,1226,1227,1228,1229,1230,1231],"menuIcon","editor-swatches","productName","requiresAllScopes","requiresLogin","requiresOrgLevelScopesOnly","title","Catalog",{"_1223":1253,"_50":1254,"_1217":1255},{"_1246":1247,"_1223":1248,"_50":1249,"_1219":1149},{"_1223":1244,"_50":1245,"_1217":1218},{"_1223":1241,"_50":1242,"_1217":1243},{"_1223":1238,"_50":1239,"_1217":1240},{"_1223":1235,"_50":1236,"_1217":1237},{"_1223":1232,"_50":1233,"_1217":1234},"Resources","/resources","misc-library","Models","/models","misc-neural-network","Helm Charts","/helm-charts","misc-ship-wheel","Containers","/containers","3d-cube","Collections","/collections","access",[1250],"Enterprise Catalog","/enterprise",[1251,1252],"NVAIE","READ","Explore Catalog","/","common-home",{"_156":157,"_158":159,"_160":161,"_162":163,"_164":165,"_166":167},{"_155":34}]